跳到正文
研踪PaperTrail

YOUR RESEARCH, IN SIGHT

让下一篇重要论文，不再错过。

覆盖计算机全学科，以 AI 为快捷入口。关注方向、作者和关键词，把发现变成持续的研究积累。

部分采集异常
正在读取账号与研究记录…

全站更新概况

+0新增论文
+0其中 AI 相关
+0已有论文修订
+100新增 AI 导读

arXiv 最近完成扫描：2026-11-09 07:12 · 全站最近入库：2026-11-08 06:02各信源进度 →

统计口径与数据新鲜度

统计 2026-11-08 08:21 — 2026-11-09 08:21（北京时间） · 点击数字查看对应简报；AI 相关包含在新增论文内。

最近完成检查：2026-11-09 07:12最近发现新增版本：2026-11-08 06:02库内论文最新更新时间：2026-11-06 02:59

每分钟刷新已保存的数据。空检查不代表采集异常；新收录可能包含历史论文。AI 导读按新增版本计数。

17,100 已收录论文14,661 AI 相关4,057 有修订112 HF 社区发现

浏览论文库

研究日报 / 周报 →按信源浏览 →

时间范围跟随排序：版本更新、本站收录或首次发表。最新收录可能包含历史论文。部分信源异常，查看状态 →

找到 5,059 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-09 07:10（北京时间）
新论文 · v1arXiv:2611.04215

Tokenizer Myths and Where To Find Them

Rylan Schaeffer, Sanmi Koyejo

原文摘要

Tokenization is a fundamental component of modern language models: before a model sees data, its tokenizer has already determined how the data appears. We identify and challenge three myths in the tokenization research literature. Myth #1: Bits-Per-Byte (BPB) is agnostic to the tokenizer and/or the vocabulary size: BPB is tokens per byte times bits per token, and the two factors do not cancel; using untrained models, we show that BPB prefers some tokenizers over others: the model has learned nothing, but the tokenizer--model pair has. Myth #2: lower BPB means a better model: On untrained models, BPB strongly prefers some vocabulary sizes, even though accuracy remains at chance, and on real models trained on the same data, BPB correlates weakly with downstream performance. Myth #3: different "tokenizer-agnostic" metrics yield consistent results: The "unigram-normalized loss" is not tokenizer-agnostic; its reported correlation of 0.989 with BPB comes from a confounder and falls to 0.057 once fixed; from the same data the two metrics select different vocabularies at each model size. We present a tokenizer trilemma: no metric can score only the true distribution best, give the same predictions the same score, and tie all untrained models, so choosing a metric is choosing a research question. We conclude with a caution for reading the tokenization literature and a conjecture for future work.

cs.CL · cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v4arXiv:2601.16836

ColorConceptBench: A Benchmark for Probabilistic Color-Concept Understanding in Text-to-Image Models

Chenxi Ruan, Yihan Hou, Yu Xiao, Guosheng Hu, Wei Zeng

原文摘要

Text-to-image (T2I) models have advanced considerably in generating high-quality images from textual descriptions. However, their ability to associate colors with concepts remains largely constrained to explicit color names or codes, while their capacity to handle \emph{implicit concepts}, such as emotions and visual states, remains underexplored. To address this gap, we introduce ColorConceptBench, an expert-annotated benchmark that systematically evaluates color-concept associations through probabilistic color distributions. ColorConceptBench moves beyond explicit color specifications by examining how models interpret 1,281 implicit color concepts, grounded in 6,584 human annotations. Our evaluation of nine leading T2I models reveals that performance varies substantially across semantic categories, and models exhibit a significant lack of sensitivity to abstract semantics. These limitations persist even when applying classifier-free guidance scaling at inference time, suggesting that achieving human-like color understanding demands a shift in how models learn and represent implicit semantic meaning.

cs.CV · cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.39562

Raincheck: Deferred Multimodal Prefill for Efficient Multi-Turn MLLM Inference

Wenjie Huang, Jun Zhang, Hongyaoxing Gu, Haoyang Huang, Tianqi Xu, Xinyue Hao, Xinyi Hu, Xinle Du 等 13 位作者

原文摘要

Multi-turn conversations with multimodal large language models (MLLMs) reuse input context, but long multimodal inputs increase time to first token (TTFT). Query-aware hard pruning reduces prefill tokens but permanently excludes information that later questions may require. We propose Raincheck, a framework for deferred multimodal prefill that decouples initial token selection from later information availability. An existing pruning method selects tokens for initial prefill; Raincheck preserves and progressively prefills unselected tokens in source order using unused compute capacity during decode. Unfinished prefill continues after answer return during available scheduling gaps. Attention visibility control and affine state mapping enable context updates after generation begins in full-attention and hybrid recurrent models. Prefill-decode fusion reduces repeated weight reads. We evaluate Qwen3-VL-8B and Qwen3.5-9B with 30% initial multimodal-token retention. Across both models and four pruning methods, Raincheck recovers on average 77.7% and 60.4% of the quality lost to matched hard pruning on MultiVerse and ConvBench, respectively. On MultiVerse, the SGLang implementation reduces mean first-turn TTFT by up to 24.1% relative to full prefill, with less than 0.5% increase in mean time per output token (TPOT).

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2509.07370

PersonaFuse: A Situation-Aware Post-Training Framework for Adaptive LLM Behavior in Enterprise Tasks

Yixuan Tang, Yi Yang, Ahmed Abbasi

原文摘要

Large language models (LLMs) are increasingly deployed in enterprise settings to support task execution and workflow automation. Despite the strong performance of general-purpose models on standard benchmarks, organizations often need to adapt them to domain-specific tasks and operational constraints in practice. However, existing post-training approaches introduce capability trade-offs and lack mechanisms for situational adaptation. In this work, we propose PersonaFuse, a post-training framework that enables situation-aware behavioral adaptation in LLMs. Grounded in Trait Activation Theory and operationalized through the Five-Factor Model, PersonaFuse dynamically adjusts model behavior based on task context using a mixture-of-experts architecture with trait-level experts and a context-aware routing mechanism. We evaluate PersonaFuse at three levels. At the capability level, it achieves balanced improvements across general reasoning, interaction quality, and safety without introducing trade-offs. At the workflow level, it improves task completion in enterprise scenarios involving tool use and structured execution across both B2B and B2C tasks. At the deployment level, we analyze model selection under cost and safety constraints using a capability-cost frontier framework, showing that models adapted with PersonaFuse achieve more favorable trade-offs and are consistently selected as optimal across deployment scenarios. PersonaFuse provides a practical framework for developing LLMs that are not only capable, but also adaptable and deployable in real-world enterprise environments.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04169

Length Is Cheap, Competition Is Not: Attention Predicts Context Rot in Retrieval

Yinuo Liu

原文摘要

Why do language models fail on long inputs? The common answer is length. We show that, for retrieval, length is cheap and competition is not. On five open models (0.5B to 3B parameters, four families), neutral filler costs nothing: accuracy stays at 0.97 to 1.00, up to 32K tokens on Qwen2.5-1.5B. Statements in the format of the evidence, placed at the same positions, drop it to 0.16 to 0.33. A single statement that gives the queried key a rival value multiplies the loss by 9 to 110, and every additional rival adds to it. Inside the model, one quantity tracks the failure: the attention mass left on the evidence. Across 126,473 trials it explains up to 58% of the variance in loss; log length explains at most 8%. The masked-filler protocol behind the claim that length alone hurts reproduces on MMLU (11 to 18 points lost at 30K tokens), but fails a visible control in retrieval: masking costs 45 accuracy points at 32K tokens, while the same filler left visible costs none. In retrieval, context rot is a competition for attention. We release the probe and all 126K trial-level measurements.

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04152

PACT: Certifying What Is Written in Parallel Decoding of Diffusion Language Models

Yanchao Li, Yusu Zhang, Jiaqing Xie, Yi Ban, Ben Gao, Zhongyu Huang, Tianfan Fu, Yuqiang Li

原文摘要

Parallel decoding speeds up masked diffusion language models by writing several tokens per forward pass. Published commit rules choose these tokens by testing the predicted distributions. Under sampling the written tokens are drawn, and these tests do not see them. We find that answer quality follows the reliability of the tokens actually written. An answer's log probability splits into this reliability and the dependence within each step. We propose PACT, a training-free commit rule that draws the tokens first, certifies the set it will write and spaces the positions of each step apart. This certificate holds at every temperature, and we show that no rule with the same guarantee commits more positions. Experiments use four models of the LLaDA and Dream families and HumanEval, MBPP, GSM8K, MATH500 and IFEval benchmarks. PACT reaches the quality of published commit rules with fewer forward passes. Spacing removes positions from each step, yet it usually lowers the total cost because later steps then write more tokens. At the same speed, published decoders also write a word twice in a row far more often than PACT. These results suggest that commit rules should certify the tokens they write. Code is available at https://github.com/LiZaiyuan0619/PACT.

cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2605.24426

SEAL: Synergistic Co-Evolution of Agents and Learning Environments

Yihao Hu, Zhihao Wen, Pan Wang, Xiujin Liu, Yitong Lin, Xin Zhang, Wei Wu

原文摘要

Large language model (LLM) agents increasingly learn through interaction, yet their training process often evolves unevenly: the policy changes with experience while the learning environment continues to provide largely fixed forms of support. As training progresses, this creates a growing mismatch between what the agent currently struggles with and what its learning process emphasizes. We call this phenomenon Agent-Learning Environment Misalignment (ALEM). We introduce SEAL, a framework for synergistic co-evolution of agents and learning environments. SEAL turns verifier-grounded execution failures into a shared diagnostic signal that coordinates two complementary forms of adaptation: it reshapes the support available in subsequent interactions and adjusts how the policy learns from collected trajectories. In this way, the learning environment follows the agent's evolving failure profile while both sides remain grounded in the same executable feedback. Across three backbones on BFCL V3, SEAL consistently improves over matched Vanilla RL, reaching 40.25% versus 30.75% with Qwen2.5-7B. A matched factorial ablation further yields a +5.00-point interaction estimate for the coupled configuration. The same trained checkpoints also show positive transfer to the evaluated BFCL V4 and $τ^2$-bench domains. These results suggest that coordinating how an agent learns with the environment it learns from is a promising direction for self-evolving tool-use agents.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04147

AdSteer: Native Advertising in Large Language Model Generation

Shuhang Lin, Chuhao Zhou, Zhenyu Wang, Xiao Lin, Changsheng Wang, Saibo Geng, Dimitris N. Metaxas

原文摘要

Large language model (LLM) services face rapidly rising serving costs. To sustain free or affordable access for a broad user base, alternative monetization mechanisms are needed. Native advertising offers a promising approach, as it creates economic value by aligning user intent with merchant offerings. Yet how to integrate sponsored content naturally into LLM responses remains largely underexplored. Existing approaches primarily rely on prompt-level injection, multi-pass rewriting, or fine-tuning, which can compromise response quality or introduce additional computational costs. We propose AdSteer, a lightweight decoding-time advertising framework that adaptively adjusts token probabilities to integrate contextually appropriate advertisement into responses. AdSteer calibrates the strength and duration of advertising influence, making sponsored content a natural part of the response. Experiments on GEM-Bench across three application scenarios show that AdSteer achieves an average improvement of 14.1% in overall score over the strongest baselines and 22.5% fewer total tokens on average than rewrite-based methods. These results demonstrate that AdSteer provides an effective and practical solution for native advertising in LLMs.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04141

Where Do Distilled Students Store Thinking?

Taeyeong Kim, Jaemin Jung, Ahhyun Kim, Unggi Lee

原文摘要

Small students degrade on large reasoning models' (LRMs') long solutions, so recent work shortens them by interpolating an LRM with a short-answer model at ratio $λ$. We ask what student weights retain when this teacher stops thinking, comparing LoRA adapters of three students from three teacher pairs at each $λ$. In each pair, the adapters' update subspace shifts most where the teacher stops thinking and solutions shorten most, locating the change without attributing it. We identify a shared adapter subspace tied to thinking-like behavior (reflective markers, monologue format and long solutions). Shared by adapters before that point, it keeps about half (post hoc) while one pair's teacher halves its solutions but keeps its think tag, and largely empties after thinking stops. Removing it largely turns this behavior off for every pair, student and benchmark. Only in that pair and one other student does it leave clearly fewer markers than removing the update's top components; elsewhere, same-energy random directions often nearly match it. Grafted into a short-answer student at twice its weight, it turns early markers (a post hoc count) on, like the whole update, but the student rarely finishes. Our work identifies a weight subspace tied to distilled thinking-like behavior.

cs.CL · cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04137

LoopKV: Loop-Aware KV Cache Compression for Serving Looped Transformers

Tianle Zhou, Yang Zhang, Tianyu Wu, Zhuohan Wang, William Nixon, Juncheng Yang

原文摘要

Looped transformers apply a shared block several times per token to improve accuracy without adding parameters. For models such as Ouro that keep a separate key-value (KV) cache per loop, cache memory limits the batch size that amortizes repeated weight reads. Uniform-bit quantization and fixed rules per loop do not allocate precision according to each layer's sensitivity in each loop. Measurements on Ouro show that quantization sensitivity varies across (loop, layer) cells, and the sum of individual cell damages approximates the overall output damage. Sharing damage also varies by cell, with later loops generally tolerating sharing better. Skipping low-attention values to save reads produces errors that compound across loops. LoopKV uses a one-time offline profile to choose 4-bit storage, 2-bit storage or reuse of the final loop's KV for each cell. The planner minimizes reads under budgets for predicted damage and storage. During decoding, each loop reads the high 2 bits of each 4-bit value and fetches the low bits only where its own attention weights require 4-bit precision. A single fused kernel in vLLM executes the plan. Serving Ouro-1.4B on one RTX PRO 6000, the memory-optimized LoopKV-M plan stores 35% less KV than uniform 4-bit quantization (KIVI-4) on the same fused kernel and CUDA graph. It sustains 1.7-2.2x KIVI-4's goodput under a latency service-level objective with a 16K-token maximum length. None of its averages on six public benchmarks falls significantly below BF16's.

cs.LG · cs.CL · cs.DC

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2606.22361

First-Token Broadcasters: Mechanistic Origins of Language Identity and Distributed Robustness in Transformers

Arjun Pillai, Christian Hoang, Anjelo Jann Laroza

原文摘要

How is language identity represented and causally maintained in transformers, and to what extent do the mechanisms identified in one language set transfer to other languages and training regimes? We introduce Language Identity Head Ablation (LIHA), a causal intervention that zeros each attention head individually and measures the resulting language switch rate. Across our multilingual evaluation sets, LIHA identifies heads whose ablation has disproportionate effects on output language. In GPT-2, a small set of first-token broadcaster heads---led by L6H1 (switch rate 0.32, 3.23$σ$ above the population mean)---attend persistently to the first prompt token across query positions. The combination of persistent first-token attention and causal sensitivity suggests that these heads participate in maintaining output language. Following ablation, first-token attention increases in heads above the ablated layer for the ablations tested (permutation test against random five-head sets, $p < 10^{-5}$). We do not test whether these heads functionally compensate for the ablated head. We also compare Qwen2.5-1.5B-Base and Qwen2.5-1.5B-Instruct, which are matched in architecture and parameter count but differ in training history. The base model is nearly flat (max SR$=$0.016, 200/336 heads at SR$=$0.0), whereas the instruct model concentrates causal influence at layer~0, led by L0H5 (SR$=$0.224, 8.93$σ$ above the mean). Finally, GPT-2 heads identified from the European-language experiments show no observed switch effects on the tested Chinese and Russian prompts, while other early-layer heads do show effects. Because script, typology, and tokenization vary together in this comparison, these results do not isolate a specifically script-driven mechanism. Repository is available at https://github.com/ajlrza/First-Token-Broadcasters.

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2609.25010

Do Synthetic Personas Predict Real Audience Response? A Headline-Ranking Study with a Generic-Reader Baseline

Alexandre Cristovão Maiorano

原文摘要

Do demographic personas improve an LLM's prediction of audience engagement? We compare a fixed ten-persona panel with a generic-reader baseline on Upworthy headlines, with supplementary news and Reddit comparisons. The primary experiment collects Gemini 3.1 Flash-Lite and GPT-4.1 observations on the same 190-package exploratory cohort, selected with simultaneous winner screening. It allocates thirty scores per arm and variant within model, retains exact raw responses, and reports metric-specific valid and paired counts. Complete-package coverage is 188 for Gemini and 188 for GPT-4.1. The baseline-minus-panel Kendall $τ_b$ gap is +0.222 (CI [+0.061, +0.385]) for Gemini and +0.257 (CI [+0.119, +0.391]) for GPT-4.1, on respectively 136 and 188 valid rank pairs. Expected top-1 is 39.1% versus 54.8% for Gemini and 42.0% versus 61.7% for GPT-4.1 (panel versus baseline). These intervals condition on the selected cohort and recorded generations and do not adjust for dependence between related experiments. The Upworthy archive documents a randomization-risk period. Excluding that period in a post-hoc sensitivity retains positive expected top-1 and lift gaps in both configurations, while Gemini tau-b and Spearman intervals include zero. Supplementary Reddit top-1 and captured lift favor the panel; these aggregate comparisons use unreconciled collection protocols and do not isolate dataset effects. A raw-replay artifact supports offline reproduction of the primary experiment. The results concern this panel, cohort, and protocol; they are not a held-out evaluation or evidence of universal persona harm. Provider request protocols differ across models, and no causal mechanism is isolated.

cs.AI · cs.CL · cs.CY

查看论文
arXiv ↗PDF ↗
版本更新 · v4arXiv:2610.11152

Imitation Dominates Reinforcement: Direct In-Context RL Is Closer to ICL Than RL

Minchan Kwon, Seunghee Koh, Sunghyun Baek, Minsung Bae, Junmo Kim

原文摘要

LLM agents increasingly improve at inference time by accumulating experience in context rather than by updating parameters. This process is often described as in-context reinforcement learning (ICRL). Whether in-context learning (ICL) can actually play the role of RL, however, has not been tested. We study this question in its simplest form, direct ICRL, where the model conditions directly on raw trajectory-reward pairs, and ask whether the reward acts as a learning signal. Through controlled experiments on four benchmarks across six models, we find that the reward is read, but its effect is small: flipping, randomizing, or removing the reward leaves the improvement curve almost unchanged, and this holds even under meta-prompts that explicitly instruct the model to explore, exploit, or reason over rewards. Trajectories drive improvement, but not through their semantic content: shuffled or corrupted trajectories work as well as real ones. These patterns closely mirror those known in ICL, suggesting that direct ICRL is better understood as a special case of ICL than as inference-time RL. This reframing has implications for agent memory design: ICL factors such as input distribution and demonstrations may matter more than RL elements such as reward shaping and exploration.

cs.LG · cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04107

Say Less, Mean Better: Training-Free Verbosity Control for Language Models

Mahmoud Alyosify, Mirna Imbabi, Eman Elkhamisi, Rahatara Ferdousi

原文摘要

Large language models (LLMs) often generate responses longer than necessary, increasing inference latency and computational cost. While prompting can encourage conciseness, controlling output length without compromising response completeness remains challenging. Aggressive early stopping can reduce generation costs but risks terminating responses prematurely. We introduce MinimalLM, a training-free framework for reliable verbosity control that combines uncertainty-aware, length-dependent stop-token adjustment with structural completeness guardrails. MinimalLM separates the decision of when to stop from whether stopping is structurally appropriate, coordinating response-level instructions, decoding-time intervention, and completion checks without modifying model parameters or requiring auxiliary models. We evaluate MinimalLM across three open-source LLMs and 330 questions, producing 5,940 responses across baseline and ablation configurations. MinimalLM reduces generated tokens by 47.6%, accelerates individual responses by 19-68%, and eliminates structurally truncated outputs in the primary evaluation. Despite substantial token savings, it achieves an average quality score of 8.04/10, compared with 8.42/10 for unmodified models and 6.46/10 for length-matched truncation. Evaluations across diverse subject domains (such as healthcare, programming, law, and science) demonstrate robust efficiency gains, while ablations establish the complementary roles of decoding control and completeness constraints. Our findings highlight that efficient generation requires more than producing fewer tokens: it requires controlling when generation ends while preserving response integrity. MinimalLM offers a lightweight, training-free approach toward this goal.

cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v4arXiv:2511.20836

Structured Prompts Improve Evaluation of Language Models

Asad Aali, Muhammad Ahmed Mohsin, Vasiliki Bikia, Arnav Singhvi, Richard Gaus, Suhana Bedi, Hejie Cui, Miguel Fuentes 等 18 位作者

原文摘要

As language models (LMs) are increasingly adopted across domains, high-quality benchmarking frameworks are essential for guiding deployment decisions. However, frameworks such as Holistic Evaluation of Language Models (HELM) typically evaluate models under a single static prompt configuration, even though model behavior depends strongly on prompt choice. As a result, reported scores can reflect prompt choice as much as model capability. Declarative prompting frameworks, such as DSPy, offer a scalable way to operationalize this evaluation protocol by evaluating models under structured prompting strategies rather than static prompts. Here, we conduct a reproducible DSPy+HELM benchmarking study and quantify how prompt choice impacts benchmark outcomes. Using five prompting methods, we evaluate four frontier and two open-source LMs across seven benchmarks against HELM baseline scores and find that prompt choice can materially impact leaderboards. Structured prompting improves performance (by 6% on average), alters comparisons (leaderboard rankings shift on $5/7$ benchmarks); chain-of-thought provides the strongest cost-performance tradeoff while more advanced optimizers offer smaller additional gains. Beyond the known intuition that prompts affect outputs, our benchmarking study integrates structured prompting into an established evaluation framework and quantifies how prompt choice alone impacts leaderboards. We open-source (i) DSPy+HELM Evaluation (https://github.com/stanford-crfm/helm/pull/3893) and (ii) Prompt Optimization Pipeline (https://github.com/StanfordMIMI/dspy-helm).

cs.CL · cs.AI · cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.18877

Lotus: Learning to Understand Time-series by Learning to Forecast

Van Dai Do, Huu Hiep Nguyen, Minh Hoang Nguyen, Manh Nguyen, Hung Le

原文摘要

Time-series language models typically align numerical signals with language using labels, captions, or question-answer pairs, without explicitly requiring their representations to preserve how the underlying signals evolve. We introduce forecast grounding, a pretraining principle that requires language-aligned time-series representations to remain predictive of the future evolution before they are adapted to downstream tasks. We instantiate this principle in Lotus, which divides each time-series channel into non-overlapping patches and maps them into the LLM embedding space. It then predicts the next patch from the LLM hidden states of the time-series tokens, encouraging the learned representation to retain temporal information. To capture both temporal changes and periodic patterns, we also propose a temporal-spectral patch representation that combines each raw time-series patch with its spectral representation before mapping it into the LLM embedding space. Across three language model backbones and six downstream evaluation datasets, Lotus achieves the best average performance with minimal trainable parameters, reaching 59.91% accuracy and outperforming the strongest matched baseline by 3.36%. Our representation analyses further show that Lotus uses the numerical input and preserves temporal structure in its learned patch representations. These results establish forecast grounding as a promising pretraining principle, enabling compact time-series language models to outperform models 25-50 times their size on TSRBench. Our code is publicly available at https://github.com/DA2I2-SLM/Lotus

cs.LG · cs.AI · cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04099

Evaluation Contracts and Evaluation Leverage: Sliced-Gap Readings in Structured Prediction Can Move Without Model Change

Hao Lu, Weixing Zhu, Ruiyang Li

原文摘要

NLP evaluation reports routinely cite sliced performance differences, most prominently the intra-inter gap in document-level relation extraction, as capability evidence. We show that such gap readings are joint properties of models and evaluation contracts: scoring function, gold version, threshold, output format, and prompt language. With predictions frozen, a method ordering reverses under a gold-standard exchange (+0.0363, CI excluding zero; tie-level after), movement localizing where gold re-partitions: renames are gauge transformations (Delta=0; re-annotation +0.862). We formalize this as an evaluation leverage account centered on the rank-flip condition. Constructively, the audited evaluation reading (AER) is a contract-marginalized, tie-declared, seed-prescribed, gate-filtered procedure whose tie declarations cover every contract-induced flip on the declared WMT19 grids (14/14; held-out 73% vs. union-of-ties 89%), matching the best single metric contract at epsilon=0.10 equivalence. Evidence spans DocRE, HacRED, a UD probe, and released gold revisions; an alternative-reference-set replication extends it beyond our testbeds. Failed predictions retained as boundaries; a 5% violation gate flags unverifiable settings and bounds survivor-selection bias. We release EvalLever, a preregistered workflow with an amendment trail; on LLM-as-judge grids, the audited ensemble stabilizes every unstable pair (13-15 to zero; generic ensemble shrinkage: the preregistered correlation criterion failed, the companion stabilization criterion passed).

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04098

When Strong Reasoners Join the Pool: Small Union Premiums and Costly Runaway Generations in LLM Routing

Teng-Ruei Chen

原文摘要

Routing benchmarks often judge a router against a hindsight oracle that counts a query as solved whenever any model's recorded answer is correct, but under stochastic decoding that oracle mixes two different objects: the clairvoyant ceiling of committing to one model per query and the premium of calling several models and keeping any correct answer. A companion study measured this premium at 3.5 percentage points (pp) for 11 small open models on MATH-500; how large it is once current reasoning models enter the pool was unknown. We measure it with a fixed-protocol, pre-committed generation on a single NVIDIA B300: all 500 MATH-500 problems, 30 seeded samples per problem, Llama-3.3-70B-Instruct and NVIDIA's Nemotron Nano 9B v2 and Nemotron Super 49B v1.5 (a NAS derivative of the same 70B) under vendor sampling, a 65,536-token cap, a pinned Math-Verify scorer, and an append-only attempt ledger, for 45,000 scored responses. The two reasoning models both reach 96.1% accuracy against 76.2% for the 70B baseline. The empirical product-coupling premium is 0.215-0.588 pp across all two- and three-model pools; the empirical plug-in selection headroom over always calling Nano is 1.27 pp; and resampling Nano three times gives practically the same empirical any-correct rate as calling all three models once. Only 31 responses reached the cap, 25 of them with exactly periodic tails, yet the capped responses account for up to 60.2% of a model's output-length-squared proxy, and long generations made a fourth planned model, Llama-3.1-8B, too slow to finish. The premium is far below the companion's 3.5 pp, but models, pool size and protocol changed together; an oracle gap should therefore be reported with the pool, the protocol, the samples per cell and the decoding budget that produced it.

cs.LG · cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04091

Same Rule, Different Verdicts: Rubric Language and Clause Structure in LLM Judges for Korean Support Dialogues

Sangdo Han, Moonsoo Kim

原文摘要

Services that operate in Korean often write their evaluation rubrics in Korean and hand them to an LLM judge. We test whether a judge's verdict depends on the language of the rubric and on how a rubric item is phrased. We write 12 criteria for support dialogues in four clause structures (affirmative, negated, conditional, exception) such that two pairs of structures have the same meaning, and we give each item in Korean and in English. Six low-cost models from three providers, plus one reference model, judge 60 translated ESConv dialogues and 144 constructed boundary cases under every combination (51,744 requests across the seven models). The analysis plan was fixed and committed before data collection. Cross-provider agreement is lower with Korean rubrics than with English rubrics (Cohen's kappa 0.687 vs. 0.745; difference -0.058, 95% CI [-0.072,-0.045]). Disagreement concentrates in negated, conditional, and exception items (odds ratio 5.6 relative to affirmative items). A single model changes its verdict between two items with the same meaning in 11-15% of cases. Two hypotheses are not supported: models from the same provider do not agree more than models from different providers, and adding paired examples does not reduce the language gap. We measure consistency only; no human labels are used, and all materials were drafted by an LLM.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04090

Swapping the Judge Changes the Ranking: A Preregistered Study of Cross-Provider Replication in LLM-Judged System Comparisons

Sangdo Han, Moonsoo Kim

原文摘要

Practitioners who compare systems with an LLM judge and have no human labels often check a conclusion by repeating it with a judge from another provider. We test how far that check can be trusted. In a preregistered study (with a disclosed post-registration reduction of the judge set), five judge models from three providers re-judged 4,000 Chatbot Arena battles (nine systems, 33 system pairs) that carry human votes, each in both response orders. Agreement with the human vote was low at the verdict level (median Cohen's kappa = 0.22), and system rankings did not survive a change of judge: Kendall's tau with the human ranking ranged from 0.17 to 0.89 (median 0.67), so the preregistered hypothesis that rankings remain stable was not supported. Conclusions replicated by two low-cost judges from different providers matched the human conclusion more often than single-judge conclusions (+0.10 precision), but the gain came from filtering out one poorly aligned judge, and replication involving that judge left a conclusion for only 35% of comparisons. Judges that agreed with humans reached a stable conclusion with no more verdicts than human votes required; the others failed to converge on 26-70% of system pairs at the sample sizes tested (up to 100 verdicts per pair). Disagreements with the human vote co-occurred across judges far more often than independence predicts, although a single human vote per battle cannot separate shared judge bias from voter noise.

cs.CL · cs.AI · cs.CY

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2502.18779

Towards Optimal Multi-draft Speculative Decoding

Zhengmian Hu, Tong Zheng, Vignesh Viswanathan, Ziyi Chen, Ryan A. Rossi, Yihan Wu, Dinesh Manocha, Heng Huang

原文摘要

Large Language Models (LLMs) have become an indispensable part of natural language processing tasks. However, autoregressive sampling has become an efficiency bottleneck. Multi-Draft Speculative Decoding (MDSD) is a recent approach where, when generating each token, a small draft model generates multiple drafts, and the target LLM verifies them in parallel, ensuring that the final output conforms to the target model distribution. The two main design choices in MDSD are the draft sampling method and the verification algorithm. For a fixed draft sampling method, the optimal acceptance rate is a solution to an optimal transport problem, but the complexity of this problem makes it difficult to solve for the optimal acceptance rate and measure the gap between existing verification algorithms and the theoretical upper bound. This paper discusses the dual of the optimal transport problem, providing a way to efficiently compute the optimal acceptance rate. For the first time, we measure the theoretical upper bound of MDSD efficiency for vocabulary sizes in the thousands and quantify the gap between existing verification algorithms and this bound. We also compare different draft sampling methods based on their optimal acceptance rates. Our results show that the draft sampling method strongly influences the optimal acceptance rate, with sampling without replacement outperforming sampling with replacement. Additionally, existing verification algorithms do not reach the theoretical upper bound for both without replacement and with replacement sampling. Our findings suggest that carefully designed draft sampling methods can potentially improve the optimal acceptance rate and enable the development of verification algorithms that closely match the theoretical upper bound.

cs.CL · cs.DS

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04079

AlgoPilot: Cross-Paradigm Reasoning in Language Models via Strategy Selection and Guidance

Yilun Hao, Ruixiao Yang, Hudson Hilal, Yongchao Chen, Kaizhi Qian, Yada Zhu, Yang Zhang, Chuchu Fan

原文摘要

Real-world problems span diverse domains, including program synthesis, symbolic reasoning, planning, and optimization, and often require fundamentally different solution paradigms. A central challenge for LLM-based reasoning is twofold: identifying an appropriate problem-solving approach and formulating it correctly. In this paper, we propose AlgoPilot, a cross-paradigm reasoning framework that enables adaptive selection of the solving approaches and structured problem formulation prior to execution. The AlgoPilot framework contains three components: 1) a SteerLM, trained via supervised fine-tuning (SFT) and reinforcement learning (RL), that adaptively selects the solving strategy (paradigm and modes) based on input problems, 2) the corresponding expert GuideLM, trained via SFT, that generates structured formulation guidance tailored to the selected strategy, and 3) an ExecutionLM that generates solutions conditioned on both the selected strategy and its formulation. Unlike prompting-based methods that rely on fixed reasoning templates, our approach learns to adaptively choose and structure problem-solving strategies based on the input. We build AlgoPilot on Qwen3-8B and evaluate across 5 problem categories and 13 benchmarks with 66 unique domains, comparing against both the latest similarly scaled small models and large LLMs. AlgoPilot significantly improves average accuracy for the same model from 38.4% to 72.9%, outperforming the best large LLM prompting baseline GLM-5 by 11.7%. Ablation studies confirm the effectiveness of each component. Notably, we show that the learned steering and formulation modules can transfer across execution models, suggesting their potential to improve other LLMs. Project page: https://sites.google.com/view/algopilot.

cs.AI · cs.CL · cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2610.04898

A Systematic Analysis of the Predictive Power of LM Surprisal in Reading Chinese

Hongao Zhu, Muxiaoqiao Xu, Yikang Liu, Siyuan Song, Yuxia Wang, Byung-Doh Oh, Hai Hu

原文摘要

This study analyzes the predictive power of LM-derived, token-level surprisal on Mandarin Chinese reading times. Using a suite of Chinese-Pythia models (14M--1.4B) trained from scratch with 30B tokens, we examine how well surprisal predicts first fixation duration, gaze duration, and total reading time in three paragraph-level eye-tracking corpora of Mandarin Chinese (GECO-CN, HKP, and MECO). Contrary to previous null findings, our results show that surprisal is predictive of Chinese reading times. However, whether predictive power scales with model size and the amount of training is corpus-specific: bigger models predict reading times better in GECO-CN, whereas inverse scaling emerges in HKP and, at the largest sizes, in MECO. Subsequently, we find that checkpoints whose surprisal remains closer to $n$-gram statistics are better predictors of reading in HKP and MECO, providing one possible explanation for the inverse scaling behavior. All in all, the predictive power of surprisal on Chinese reading time measurements is corpus-specific, which cautions against drawing conclusions about language model scale from a single corpus.

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04031

Language Model Augmented Decision Transformer for Multi-Task Offline Reinforcement Learning

Xinbo Zhao, Yuhang Liu, Xin Zhang, Lei Le, Zhaohan Xi, Yu Yang, Yingxue Zhang

原文摘要

Offline multi-task reinforcement learning (MTRL) aims to learn a unified policy from heterogeneous offline datasets that generalizes across tasks. However, existing offline MTRL methods often struggle with two fundamental issues: severe per-task data scarcity and limited cross-task generalization. In this paper, we propose Language-Model Augmented Decision Transformer (LaMA-DT), a unified sequence-modeling framework that leverages a single large language model (LLM) for both trajectory augmentation and multi-task policy learning. LaMA-DT uses natural-language task descriptions to guide task-aligned trajectory augmentation through a set of supervised generation and alignment objectives, and incorporates a scoring-and-filtering mechanism to retain high-quality augmented trajectories. The same LLM is then repurposed as a language-conditioned Decision Transformer and fine-tuned on a mixture of real and augmented data. Experiments on Meta-World and MuJoCo multi-task benchmarks demonstrate that LaMA-DT consistently outperforms competitive baselines, particularly in low-data regimes, showing that language-guided data augmentation substantially improves cross-task generalization.

cs.CL · cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.04003

WASP-Mem: Let Agent Organize Procedural Memory. Human Spend Effort on Diagnosis

Peizhi Niu, Ge Shi, Vaibhav Gorde, Sarthak Ghosh, Christos Faloutsos, Min Cheng, Kevin Chen

原文摘要

LLM agents struggle to automate complex enterprise procedures because static, pre-built harness cannot capture proprietary policies, undocumented behaviors, or state-dependent recovery paths that surface only during execution. Procedural memory accrued from agent experience offers a training-free path to closing this gap -- yet unlike episodic records or semantic facts, it must consolidate repeated successes and diagnosed failures into reusable actions and recovery strategies. We introduce WASP-Mem (Wiki-based Agent-Synthesized Procedural Memory), a nimble but effective method that distills execution traces and feedback into persistent, interlinked wiki pages, enabling continual agent learning. On TheAgentCompany benchmark, WASP-Mem raises the checkpoint pass rate by 4.6-6.1% over the strongest baseline we compared across three models while requiring 57-61% less storage. We further present a systematic controlled study spanning consolidation method, feedback type, trace ingestion scope, and refinement rounds. Our results reveal that prescriptive knowledge organization such as human defined schemas and explicit deduplication rules hurts performance rather than closing the procedural-knowledge gap. In contrast, human diagnosis converts failed traces into corrective knowledge and drives sustained improvement: a single round of human diagnosis outperforms four rounds of self-judgment by 7.0-10.9%. These findings point to a simple division of labor: let agents organize procedural memory, and invest human effort where it matters most -- diagnosis.

cs.AI · cs.CL · cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03998

Dory1.0-4B: A Looped Normalized-GPT

Maxim Fishman, Simeng Sun, Shantanu Acharya, Faisal Ladhak, Brian Chmiel, Santiago Akle Serano, Ron Banner, Boris Ginsburg 等 9 位作者

原文摘要

We introduce Dory1.0-4B, a 4-billion-parameter looped (latent-recurrent) language model with elastic depth: a single checkpoint serves different inference compute budgets. Dory1.0-4B is built on a normalized Transformer (nGPT) backbone with interleaved sliding-window and full-attention layers, organized into an input block, a shared recurrent block, and an output block. At inference, the recurrent block can be applied a variable number of times. We pre-trained Dory1.0-4B on 6 trillion tokens. The first 4.5 trillion tokens were trained in NVFP4, and additional pre-training then extended the context length to 256K tokens. Next, we post-trained the model using Supervised Fine-Tuning (SFT) and Reinforcement Learning with Verifiable Rewards (RLVR), with a difficulty-aware loop curriculum that routes harder prompts to deeper loops. Increasing the loop count improves accuracy, with the largest gains on coding, agentic, and long-context benchmarks. With two or more loops, Dory1.0-4B outperforms larger models such as Qwen3.5-9B, Gemma 4 12B, and Gemma 4 26B-A4B on average across 12 benchmarks. Looping also makes Dory1.0-4B more token-efficient on agentic benchmarks: with four loops, it generates 1.3-2.5x fewer tokens than with one loop, as the extra computation is spent in latent space rather than on longer outputs. To reduce memory during looping, Dory1.0-4B caches keys and values only from the last loop, keeping its KV-cache footprint equal to that of a non-looped model at any loop count while retaining most of the accuracy gains. We open-source the BF16, MXFP8, NVFP4, and GGUF quantized checkpoints.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03993

Whose startup does the AI back? An algorithm audit of reputation and demographic signals in LLM-assisted crowdfunding

Mirza Samad Ahmed Baig, Syeda Anshrah Gillani

原文摘要

Large language model (LLM) assistants increasingly advise backers on which crowdfunding campaigns to support. LLMs have known race and gender problems, and it is unsettled whether they reproduce, erase, or reverse the discrimination documented in entrepreneurial finance. We ask when, and through which decision process, LLM assistants use founders' gender and ethnicity. In a pre-specified choice-based conjoint experiment, seven LLMs chose among five synthetic campaigns with independently randomized quality signals and founder names, under three backer personas and a persona-free control (40,248 responses). By default, founder names barely matter: under budget-conscious and delivery-focused personas, gender and Black-name effects are statistically equivalent to zero. When the backer wants to support underrepresented creators, the models favor Black-signaled (+8.1 percentage points; positive in all seven models) and female-signaled (+5.1) founders. They do so most strongly among near-tied campaigns and for the founder who alone carries the signal, and they almost never name gender or ethnicity in their reasons. The results fit an agency account of algorithmic discrimination: an aligned assistant takes its demographic preferences from its user rather than from its training data.

cs.CL · cs.AI · cs.CY · cs.ET

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03992

FlagRot: Selective Rotation for Low-Bit KV Caches

Kunming Shao, Xijie Huang, Jierun Chen, Jiaqi Shao, Yanli Wang, Jiangnan Yu, Qiwei Li, Wei Zhang 等 10 位作者

原文摘要

Long-context inference is bound by the KV cache, which 2-bit codes could shrink about sixfold if their error stays in check. The usual remedy rotates keys before quantizing, assuming rotation always helps. However, we observe that it does not: rotation flattens a spiky piece of a key but can give a flat one a new peak, and both coexist within one key. We therefore propose FlagRot, which stores each 32-channel piece rotated or as is, whichever has the smaller calibrated attention error, and rotates the query instead when reading. We further show that how often to decide follows the quantization axis: per-channel caches such as KIVI need one decision per head, at no write cost, while per-token caches decide per key, with the choice stored in existing sign bits. At 2.5 bits per value, FlagRot lowers the twelve-model KL divergence of KIVI from 18.7 to 12.4 x 10^-3, below RotateKV on 9 of 12 models. Served as a vLLM plugin, it raises RULER at 128k tokens by +2.9 points and the LongBench v1 mean by +0.4 to +1.9 points on the three of four models where 2-bit KIVI loses accuracy. It also lowers the error of 4-bit integer and FP4 caches.

cs.CL · cs.AI · cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03985

Thinking Fast, Trading Late: System One Decision Models vs the Market

Amir Dehkordi, Alireza Kargarzadeh, Nariman Khaledian, Navid Parvini, Sid Ghatak, Arman Khaledian

原文摘要

Large language models (LLMs) read company news at scale, and backtests suggest that their readings predict stock returns. We ask how much of that predictability survives a test that rules out look-ahead, memorization and selection. We propose OLIVIA, a protocol for such tests: point-in-time inputs with company identities masked, selection on validation data only, an analysis plan frozen by cryptographic hash, and a test period unblinded once through a logged lock. We apply it to AI decision models, non-generative language models that return probabilities for typed questions, reading news about S&P 500 stocks: an open-weights model we fine-tuned on 2024-2025 data, three commercial models and standard baselines. The fine-tuned model's signal replicated on two validation quarters (daily information coefficients 0.055 and 0.065) but fell to 0.020 (t = 1.7) on the blinded January-July 2026 test. The commercial models' readings line up with the overnight move to the open (information coefficient 0.074 to 0.077) but not with the move after it, so the news is priced by the time a trader can act on it at the open. No model can tell over- from under-reaction, and no pre-registered strategy survives costs and a deflated Sharpe ratio. Commercial models identify masked companies in 62% to 74% of cases against a 28% baseline: masking hides names, not companies. The market reads the news first.

cs.AI · cs.CL · cs.LG · q-fin.ST

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.03982

SLATE: A Large-Scale Benchmark for Long-Horizon API Selection in E-commerce Tool Use

Rongzhe Wei, Ge Shi, Min Cheng, Sarthak Ghosh, Vaibhav Gorde, Leman Akoglu, Pan Li, Na Zhang

原文摘要

Enterprises increasingly want LLM agents to automate standard operating procedures over their software stacks -- mapping each step of a procedure to the right call among thousands of internal APIs, without hand-authored step-to-API mappings. The core difficulty is API selection at scale: choosing and composing the correct tools over a long horizon, judged only by whether the overall task succeeds. Current tool-use benchmarks do not test this regime -- they expose only a handful of tools, evaluate single-step invocation, or grade multi-step behavior with subjective LLM-as-a-judge scores that penalize valid alternative trajectories. We introduce SLATE (Synthetic Large-scale API Toolkit for E-commerce), a large-scale benchmark that tests large-toolset, long-horizon API selection in a realistic enterprise domain. SLATE couples (i) queries decomposed into hierarchical, tool-grounded procedures over a ~1,000-tool e-commerce library, (ii) a deterministic, context-aware simulator that returns recorded outputs for matching calls and default responses otherwise, and (iii) an end-to-end task-success metric -- a task succeeds when it reaches every reference outcome -- that credits diverse-but-valid trajectories while remaining fully automatic and reproducible. Experiments with six open decision models and the language models they are built on, from 3B to 12B parameters, show that even the best model completes only 5.0% of the tasks, establishing SLATE as a challenging testbed for reliable enterprise tool use. We release SLATE, its simulator, and the evaluation harness at https://geshijoker.github.io/slate/

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗