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 09:43 — 2026-11-09 09:43（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:2611.02480

Chain on Demand: Deferred Evidence Organization for Long-Term Conversational Memory

Di Lu, Zechen Li, Hao Xue

原文摘要

Long-term conversational memory faces a timing mismatch: write-time summaries and graphs organize history before a future query is known, while independent retrieval can allocate a limited evidence budget to redundant or fragmented turns. We introduce Chain on Demand (CoD), a query-time evidence organization framework over an unchanged turn-level memory. Given a broad candidate pool, CoD constructs a temporary evidence chain by allocating a fixed anchor budget according to query relevance, lexical diversity relative to selected anchors, and newly covered query terms, followed by deterministic restoration of each selected anchor's same-session neighborhood. The resulting chain is used only for the current query, while the underlying conversational memory remains unchanged. In a paired comparison over all 1,986 LoCoMo questions, CoD achieves 82.2% accuracy on 1,540 non-adversarial questions and 76.3% on all 1,986 questions. On LongMemEval-S, one fixed configuration that indexes both user and assistant turns obtains 86.6% across all 500 questions. These results show that long-term memory can preserve the original conversation while deferring evidence organization until the information need is known.

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.40726

Li$^{2}$SAO: Lifting Light Single-Rollout Asynchronous RL Optimization via Group-Relative Actor-Critic Bootstrapping

Zichen Bai, Yanli Wang, Yang Liu, Junhao Shi, Xueqin Liu, Jinguo Sun, Muye Li, Shaoqiang Zheng 等 12 位作者

原文摘要

Asynchronous reinforcement learning (RL) is becoming a mainstream paradigm for post-training large language models (LLMs), as it improves training efficiency by updating the model as rollouts arrive. Recent work argues that group-wise sampling in GRPO does not naturally fit asynchronous RL training because rollout lengths follow a long-tailed distribution, and proposes Single-rollout Asynchronous Optimization (SAO) to improve stability and off-policy behavior, especially on agentic tasks. However, SAO relies on a closed-source value model obtained from scaling value pretraining. In this paper, we analyze GRPO and LiSAO, our lightweight reproduction of SAO without the pretrained value model, on single-turn reasoning and multi-turn agentic tasks. We find that GRPO outperforms LiSAO on reasoning tasks, while the reverse holds on agentic tasks. Since the value model of LiSAO is not yet well trained , LiSAO lags behind GRPO substantially in early training steps. To address this issue, we propose Li$^{2}$SAO, which Lifts Light Single-rollout Asynchronous Optimization with group-relative actor-critic bootstrapping: while the critic is cold, a group-relative advantage trains the actor and a group-centered return trains the critic, after which training switches to single-rollout GAE. Experiments show that Li$^{2}$SAO catches up with or surpasses GRPO in early training and further raises the later performance ceiling. Code is available at https://github.com/jd-cc-coder/Li2SAO.

cs.CL · cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2609.34428

AgentHop: A Diagnostic Benchmark for Agentic Multi-Hop Question Answering over Scientific Citation Graphs

Chanhee Park, Jeongho Yoon, Sungbin Han, Hyeonseok Moon, Heuiseok Lim

原文摘要

Agentic tasks require a large language model to interact with the world, navigating information and gathering evidence across multiple steps with restricted resources. Due to this complexity, agentic task failures arise from various sources, and pinpointing these failure causes is essential to diagnose and improve agentic systems. Existing benchmarks, however, tend to focus on a single leaderboard score, leaving the underlying failure modes opaque. To fill this gap, we introduce AgentHop, a diagnostic benchmark of 1,011 multiple-choice questions paired with a controlled seven-tool sandbox under fixed token, turn, and tool-call constraints. AgentHop reveals model vulnerabilities by dissecting a single accuracy score along four axes of agent operation: retrieval, synthesis, tool-call, and resource management. Across 19 models, we find that behavior clusters by model family, with tool-call signatures revealing distinct family fingerprints: GPT models commit early, Anthropic and GLM checkpoints verify before committing, DeepSeek and Kimi over-search, and Gemini-3 Pro stays balanced. Decomposed axes further expose within-family structure: Claude Opus 4.6 and Sonnet 4.6 land within one accuracy point yet diverge on retrieval-versus-synthesis emphasis, with Opus retrieving more and Sonnet synthesizing better. We release the full benchmark set and the harness to support diagnostic agent benchmarking.

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2606.08239

When No Answer Is Correct: Diagnosing Absent Answer Detection for MLLMs in Video Understanding

Yiheng Wang, Yueqian Lin, Lichen Zhu, Yudong Liu, Hai "Helen" Li, Yiran Chen

原文摘要

Multimodal large language models (MLLMs) have made substantial advancements in video understanding, yet the reliability of their responses remains underexplored. This work presents a diagnostic study of absent answer detection for MLLMs in video understanding, where the correct answer is deliberately excluded from the candidate set and a reliable model is expected to recognize that no valid option exists. We evaluate the absent answer detection behavior under three settings: multiple-choice questions augmented with a ``None of the Above'' option, open-ended generation with a detection instruction, and standard evaluation without any guidance. Across a diverse set of models and benchmarks, we find that MLLMs overwhelmingly select plausible distractors rather than detecting the absent answer. This failure is more pronounced in temporal reasoning tasks and worsens with denser frame sampling. We further explore chain-of-thought prompting as a mitigation strategy and find that while it substantially improves detection rates, performance remains unsatisfactory, suggesting that prompting-based strategies alone are insufficient to fully address this limitation. These findings expose a systematic failure in absent answer detection and highlight the need for explicit detection mechanisms in multimodal systems.

cs.AI · cs.CL · cs.CV

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.23344

Language Models Learn to Improve Language Models

Minhua Lin, Zijun Wang, Zewen Liu, Zhaotian Weng, Bing He, Zhan Shi, Benoit Dumoulin, Xiang Zhang 等 10 位作者

原文摘要

Large language models (LLMs) are increasingly used to automate AI research, including language-model (LM) development, offering a potential path toward recursive self-improvement. However, turning experimental feedback into a learned ability to propose effective improvements across models faces two challenges. First, training such an ability with real feedback is expensive: reinforcement learning (RL) requires evaluating tens of thousands of proposals, each potentially requiring a full model-training run. Even at ten GPU-hours per run, reward collection alone would consume hundreds of thousands of GPU-hours, motivating a cheaper surrogate. Second, surrogate feedback can become unreliable as policy optimization favors overestimated proposals when the proposal distribution shifts. To address these challenges, we propose LM2LM, which trains LLMs with RL to generate effective LM improvement proposals. To provide scalable feedback, we use a research world model (RWM), a frozen LLM that predicts proposal gains from real experimental records provided in context, as the reward model for RL. To maintain feedback reliability, we periodically execute current-policy proposals ranked highly by the RWM and add their measured outcomes to its context. Across eight autoregressive and diffusion LM development environments, real-execution evaluations on held-out tasks show that LM2LM enables Qwen3.8-27B to outperform GPT-6-sol, even when GPT-6-sol is given 300 real experimental records in context, while requiring 33% lower data-generation cost. LM2LM also generalizes to a model family unseen during training.

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02438

As Expected? Efficient LLM Planning via Expectation-Observation Consistency Verification

Qingfu Zhu, Huanyang Zheng, Ting Liu, Wanxiang Che

原文摘要

LLM planning often relies on reactive executors (e.g., ReAct) that generate a single action at each step to remain grounded in the latest observations, incurring substantial inference cost. Batched executors reduce this cost by predicting multiple actions in a single model invocation, yet they are prone to compounding errors, as intermediate observations are ignored. We propose an expectation-observation consistency verification (EOCV) mechanism for batched executors, preserving grounding while improving efficiency. The mechanism generates expectations of action outcomes and compares them with the resulting observations. Consistency permits continued execution, whereas mismatches trigger early termination and replanning from the current observation. Experiments on text-based and multimodal benchmarks across multiple backbones show broadly comparable average task performance between EOCV and reactive executors, with a 29-35% reduction in average weighted token usage. Moreover, the mechanism is model-agnostic and remains effective when integrated into state-of-the-art planners.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02432

Diagnosing Model--Item Interactions in Language Model Benchmarks with LSIRM

Jinwen Luo, Han Jiang, Ziang Xiao, Susu Zhang

原文摘要

Language models with similar benchmark accuracy can succeed on different items. Overall scores do not show where their performance differs or which item properties might explain those differences. We demonstrate how the latent space item response model (LSIRM) can help evaluators investigate these patterns and their consequences for model comparison. We analyze 1,348 language models on 5,761 BBH items and compare ways of presenting and scoring answers across BBH, GPQA, MMLU-Pro, and MuSR. We examine LSIRM maps, which summarize differences in performance across models and items, alongside item properties and records of correct and incorrect responses. In BBH, nearby items on these maps often share the same correct answer, such as "Yes" or "No." Across the four benchmarks, this pattern is stronger when answers are scored using option labels, such as A, B, C, and D, than when using the text of the options. We then estimate each model's tendency to perform better or worse on items with particular correct answers and quantify how this tendency contributes to its score. This separates an estimated contribution associated with correct answers from the remaining differences in item performance. Two models' total numbers of correct answers differed by only 14 out of 5,761 items. Excluding the estimated answer-category contribution from their scores moved their ranks hundreds of places in opposite directions, while keeping all other estimates unchanged. The approach helps evaluators identify differences hidden by overall accuracy and understand how the choice of what to include in a score affects model rankings.

cs.CL · cs.LG · stat.ME

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02430

InteractionBench: Measuring Timing, Task Accuracy, Whisper and Volume, and Grounding in Real-Time Voice Systems

Richard Yucheng He, Chen Xu, Yihang Liu, Tairan Chen, Baodong Cao

原文摘要

Most benchmarks for real-time voice systems focus on one aspect, such as turn-taking, instruction following, paralinguistic control, or conversational memory, each with its own data and set of systems. We present InteractionBench, a benchmark and evaluation harness that measures four dimensions on the same systems from the same recorded user audio: conversational timing, spoken-task accuracy, whisper and volume, and groundedness in conversations of more than twenty turns. The benchmark contains 221 items per system in thirteen scenarios. Before any system runs, we fix the expected floor-control actions for spoken tasks by rule and take the expected answers from scripted dialogues. We score whispering from the audio signal. We evaluate GPT-Live, GPT-Realtime, Gemini Live, Moshi, PersonaPlex, Freeze-Omni, and a turn-based cascade in 1,519 graded runs. Among the six full-duplex systems, timing and spoken-task accuracy are rank-correlated ($ρ= 0.83$), while the cascade has the highest spoken-task score (38/50) and the second-lowest timing score. No system corrects a user's grammar within the required window. Only Gemini Live whispers on most requests (8/10). Hosted systems pass 67-88% of groundedness probes, compared with 14-35% for open full-duplex models. Timing is measured where the client hears each system, so the timing of hosted systems includes network delay: for GPT-Live, whose server reports its own clock, removing a delay of about 1.5 s lowers its timing score from 86% to 56%. We also document where the protocol differs across systems, including the text instructions.

cs.SD · cs.CL · eess.AS

查看论文
arXiv ↗PDF ↗
版本更新 · v5arXiv:2509.24773

ViSSFlow: Unifying Video-Conditioned Sound and Speech Generation via Joint Learning

Xin Cheng, Yuyue Wang, Xihua Wang, Yihan Wu, Kaisi Guan, Yijing Chen, Peng Zhang, Xiaojiang Liu 等 10 位作者

原文摘要

Video-conditioned audio generation, including Video-to-Sound (V2S) and Visual Text-to-Speech (VisualTTS), has traditionally been studied as two separate tasks despite their shared objective of generating audio conditioned on visual content. Existing methods rely on task-specific architectures and training paradigms, while the possibility of a unified generative framework remains largely unexplored due to heterogeneous conditioning signals and limited joint training data. In this paper, we propose ViSSFlow, a unified flow-matching framework for both V2S and VisualTTS. To effectively integrate diverse conditioning modalities within a Diffusion Transformer (DiT), we introduce a disentangled condition aggregation mechanism that exploits the complementary roles of attention layers, using cross-attention to model semantic conditions and self-attention to capture temporally dense visual cues. Furthermore, contrary to the common assumption that joint optimization of V2S and VisualTTS causes task interference, we show that a single model can be trained end-to-end without sacrificing task-specific performance. To address the scarcity of paired training data covering both sound and speech generation, we further propose a simple yet effective feature-level data synthesis strategy that enables robust joint learning. Extensive experiments on V2S, VisualTTS, and joint generation benchmarks demonstrate that ViSSFlow consistently outperforms state-of-the-art task-specific baselines while providing a unified solution for video-conditioned audio generation. Code and demos are provided in Supplementary Material.

eess.AS · cs.AI · cs.CL · cs.CV · cs.SD

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02404

When Rewards Depend on the Group, GRPO Plays a Tournament

Junyi Gao, Yu Shi, Pingzhao Hu, Ewen M Harrison

原文摘要

Rewards for diversity, coverage and distribution matching make the reward of each sampled response depend on the rest of its group, and they are now routinely optimized with GRPO and RLOO. We prove that in this setting the group-relative baseline changes the objective. The expected update equals each response's own reward gradient plus $α$ times its effect on its groupmates' rewards, with $α=-1/(K-1)$ for groups of size $K$, so GRPO trains every response to beat its groupmates in a zero-sum tournament. It cannot see a pairwise diversity bonus when K=2, sees less than half of it at any K, and often does worse than using no baseline at all. Setting $α=+1$ gives Team-GRPO, an unbiased estimator for any group-dependent reward that works with two samples per prompt and requires no extra model calls for pairwise rewards. The theory predicts training quantitatively. It locates the fixed points of a 9B policy to within 0.02 in $\ell_1$ and predicts how concentrated trained policies become with Pearson r=0.995. On GlobalOpinionQA, GRPO collapses onto the majority answer and ends further from human opinion than the untrained model, while Team-GRPO cuts the error from $0.20$ to $0.08$. On NoveltyBench, Team-GRPO avoids the diversity collapse of GRPO and significantly improves set utility.

cs.LG · cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2607.05849

CoPiT: Cognitive Pivot Translation for Digraphic Low-Resource Mongolian in the Traditional Script

Burte Bayarsaikhan, Serynn Kim, Buru Chang

原文摘要

Low-resource languages remain challenging for machine translation, and Mongolian is a representative case. As a digraphic language, Mongolian is written in both Cyrillic and Traditional scripts, which exhibit a severe imbalance in data availability. While the Cyrillic script is relatively well-resourced, the Traditional script remains extremely data-scarce and orthographically ambiguous, leading to substantial performance degradation in direct translation. We propose CoPiT, a cognitively motivated pivot-based translation pipeline that exploits this internal resource hierarchy by routing translation through the Cyrillic script. The pipeline explicitly resolves script-induced ambiguity in the Traditional script before translation, enabling more stable and accurate meaning transfer. Across multiple backbone models and target languages, CoPiT generally improves over direct translation, with substantial absolute BLEU improvements and up to 1.5-1.6x gains in COMET. These gains allow strong open-source models to match or outperform GPT-4.1 under comparable evaluation settings. Beyond inference-time improvements, CoPiT enables the construction of synthetic parallel data directly from Traditional-script text, mitigating data scarcity in realistic low-resource scenarios. We release a new multi-script parallel dataset covering Mongolian in both scripts alongside English, Korean, and Russian.

cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2606.23321

Tmax: A simple recipe for terminal agents

Hamish Ivison, Junjie Oscar Yin, Rulin Shao, Teng Xiao, Nathan Lambert, Hannaneh Hajishirzi

原文摘要

Terminal-using agents have quickly become a popular and important downstream application of language models (LMs). Despite their prevalence, relatively little academic work has examined RL-based training of these models, due to difficult benchmarks, a lack of data, and a lack of simple baseline recipes. We present TMax, an open RL recipe for terminal agents that brings small open models closer to the frontier. While simple, our recipe achieves 33% on Terminal-Bench 2.0 with only 9B parameters, outperforming much larger models from prior work. Our recipe first involves generating data using a novel taxonomy, combining difficulty control, personas, and verifier diversification, which allows us to cheaply generate large amounts of terminal environments for RL and SFT training. We then train open-weight models using RL with our data, using a simple, outcome-only recipe, with small adjustments to improve stability. Our recipe works for models from 2B to 27B and across model families, dominating the Pareto curve for open terminal-agent performance at smaller sizes. We release our data, models, and code as a strong baseline for future open academic work on terminal agents at https://github.com/hamishivi/tmax.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02395

Solver-Like Behavior in Large Language Models as Formalizers

Rikhil Amonkar, Stuti Mohan, Kevin Nguyen, Ceyhun Efe Kayan, Ronan Le Bras, Li Zhang

原文摘要

In the LLM-as-formalizer paradigm, a large language model translates a problem into a formal program and leaves the search to a solver, gaining the solver's guarantees. Formalizers are now often reasoning models that write a long reasoning trace before the program, yet they are judged only by the program's output. We find that the trace routinely does the solver's job, assigning, testing and ruling out values until it often reaches the full answer; we call this solver-like behavior (SLB). We define an ontology of SLB over the trace, the program and its output, and use LLM judges, checked against human labels, to label 2,000 generations from three reasoning models on four constraint satisfaction datasets, each formalized in a declarative and an imperative formal language. SLB appears in 85% of traces and takes 24% of reasoning tokens, yet the program usually recomputes the answer, so the effort is redundant. The 9% of programs that inherit the trace's answer also inherit its errors, and pass output-based evaluation whenever that answer is correct. The amount of SLB is not associated with accuracy, and SLB is more common in imperative Python than in declarative Z3 or PDDL. Cautionary prompts reduce it by only a quarter, but hiding the problem's values behind placeholders, which the program receives only at run time, removes 97% of it without loss of accuracy. Formalizers should therefore be evaluated by how their programs are reached, not only by what they output.

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2605.11442

Mobius Injection: From One-Shot Prompt Injection to Sustained Resource Exhaustion in AI Agent Infrastructure

Zi Liang, Ronghua Li, Yanyun Wang, Qingqing Ye, Haibo Hu

原文摘要

Large Language Model (LLM) agents have emerged as key intermediaries, orchestrating complex interactions between human users and a wide range of digital services and LLM infrastructures. While prior research has extensively examined the security of LLMs and agents in isolation, the systemic risk of the agent acting as a disruptive "hub" within the user-agent-service chain remains largely overlooked. In this work, we expose a novel threat paradigm by introducing Mobius Injection, a sophisticated attack that weaponizes autonomous agents into zombie nodes to launch what we define as Agent-based and -Oriented DDoS (AbO-DDoS) attacks. Mobius Injection composes known indirect prompt injection and persistent agent components through a distinct structural mechanism named Semantic Closure, which induces sustained recursive execution of agent components through a single injection event. We demonstrate that this attack is lightweight for the adversary, can evade traditional volumetric DDoS monitors and input filters calibrated for low false-positive rates, and is highly configurable, allowing for surgical targeting of specific environments or model providers. To evaluate the real-world impact, we conduct extensive experiments across three representative claw-style agents and three mainstream coding agents, integrated with 12 frontier proprietary or open-weight LLMs. Our results demonstrate that Mobius Injection achieves substantial attack success across diverse tasks, driving single-node call amplification up to 51.0x. Under coordinated multi-node pressure, each infected node still drains about 123k--127k tokens on a shared five-GPU backend; benign request latency rises by abou 23x--24x, and coordinated infections eventually saturate shared capacity until benign traffic is effectively denied service.

cs.CR · cs.AI · cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02387

EviCast: Forecasting Retrieval Gains Through Shared Response Calibration

Yibo Wang, Jisheng Dang, Bimei Wang, Jiongpei Hou, Jingwen Zhao, Wenjun Ma, Fei Shen, Juncheng Li 等 12 位作者

原文摘要

Retrieval-augmented generation (RAG) extends large language models with external evidence, yet relevant evidence does not ensure that a generator can use it. Indiscriminate retrieval can degrade task performance while adding retrieval and long-context inference costs. Existing methods neither determine before large-scale execution whether a generator, task distribution, and retriever configuration is worth deploying nor diagnose multiple candidate retrievers under a shared budget. We introduce EviCast, a framework that forecasts retrieval value through interventional evidence-state decomposition. EviCast measures a generator's response to controlled evidence states once and combines this shared profile with the offline evidence-state distribution of each candidate retriever, yielding the end-to-end net utility and its 95% lower confidence bound without executing every RAG configuration. A state-refinement analysis characterizes when its binary and topology-aware instantiations are exact and when shared calibration is statistically more efficient than per-retriever calibration. In frozen evaluations, EviCast correctly decides all 14 HotpotQA retrieval configurations with a forecast error of 2.75 F1 points while reducing diagnostic generation calls by 80%, and it attains the lowest forecast error among seven predictors in the low-budget regime. Its topology-aware instantiation, EviCast-T, captures the complementarity between terminal and bridge evidence, correctly deciding all 16 configurations on MuSiQue and HotpotQA and 23 of 24 after prospective transfer to a third dataset. These results show that retrieval deployment depends on both evidence availability and the generator's ability to convert evidence structure into stable net utility. Code is available at https://github.com/YiboWang2002/EviCast.

cs.CL · cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2606.11806

External Experience Serving in Production LLM Systems: A Deployment-Oriented Study of Quality-Cost Trade-offs

Lin Sun, Heming Zhang

原文摘要

Production LLM systems accumulate reusable operational experience, but the practical deployment issue is not merely whether such experience can help. It is how different serving strategies trade off quality against online cost under realistic constraints. Injecting external experience can improve task quality, yet it also increases prompt burden, latency, and serving pressure. We study \textit{external experience serving} as a deployment-oriented quality-cost trade-off problem. We evaluate this question in a real production moderation setting, with tool-use and GPQA as supporting contrast tasks that expose different output-cost regimes. We compare no-experience baselines, random experience controls, global prompt injection, and retrieval-based selective injection, and analyze both task quality and serving cost. The results show that, once experience becomes case-dependent, selective retrieval provides a stronger operating point than unconditional global injection. They further show that retrieval quality matters more than simply increasing Top-$K$, and that the same serving policy can exhibit substantially different cost-benefit profiles across short-output and decode-heavy regimes. These findings suggest that external experience is best treated as a selective, cost-aware serving decision rather than as a universal add-on. Overall, in the settings studied here, external experience pays off only when both the serving interface and the task-specific cost structure make its quality gains worth the online cost.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02368

Beyond Accuracy: How Quantization Affects Coding Agents

Areej Mehboob, Muhammad Ali Shafique, Muhammad Usman Qadeer, Hamza Farooq

原文摘要

Agents are increasingly used for coding tasks, but the inference cost of large language models (LLMs) limits where they can run. Quantization reduces model size, but its effects on coding agents that execute and revise programs remain underexplored. We compare BF16 with 4-bit AWQ and GPTQ for Qwen2.5-Coder-Instruct (7B, 14B, and 32B) and Devstral-Small-2507 (24B) on OJBench coding problems across three seeds. We detect no statistically significant accuracy loss for any of the coding agents. However, final accuracy can conceal substantial behavioral changes: AWQ reduces Qwen-7B one-shot solves from 17 to 0 while increasing loop-based solves from 10 to 25. All Qwen-32B successes involve tool use, whereas Devstral rarely makes a valid executor call. Failure profiles also differ: Qwen-32B produces far more compilation errors than the smaller Qwen models, rising slightly under quantization, while Devstral's failures shift significantly from runtime errors toward wrong answers and timeouts. These findings indicate that quantization robustness in coding agents should be assessed through solution trajectories, tool use, and failure modes, in addition to final accuracy.

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02353

VisCodeBench: a new benchmark dataset for Text-to-Vis reflecting real-world practice

Hy Nguyen, Thomas James Elton, Ella Wilson, Andrew Reeson, Cecile Paris, Josiah Poon, Jonathan K. Kummerfeld

原文摘要

Generating visualisations from natural language requests (text-to-vis) enables users to explore data more efficiently. While LLMs can complete simple visualisation requests, their performance on challenging ones is not well covered by existing benchmarks. We introduce VisCodeBench, an end-to-end text-to-vis benchmark with 2,672 samples from three different domains: educational, scientific, and general audience scenarios with varying levels of complexity. We benchmark three visualisation systems---Codex, Claude Code, and LIDA---paired with a range of open and commercial LLMs across seven metrics covering execution, chart content, and visual quality. While basic execution is near-solved for strong systems, no system comes close to reproducing the reference data, with the best displayed data similarity reaching only $0.81$ and falling to $0.69$ on our hardest subset for this metric. We further find that frontier models separate on chart correctness and efficiency, with ClaudeCode-Opus-4.7 improving chart correctness by around 5 points over Codex-GPT-5 while running roughly 20 seconds faster. Separately, we find that system design matters as much as the underlying model: despite both running on GPT-5, Codex clearly outperforms LIDA, showing that the surrounding prompting, execution, and refinement pipeline can shape outcomes as much as the model itself. Finally, latency remains a practical obstacle, as even the strongest systems take 39-78 seconds to produce results. VisCodeBench is available at \href{https://github.com/SAIPLL/VisCodeBench}{https://github.com/SAIPLL/VisCodeBench} and accessed via a leaderboard; the full dataset is kept private to prevent LLM training use.

cs.CL · cs.HC

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02348

Which Words a Model Fears: From a Refusal Direction to a Vocabulary-Wide Fear Lexicon

Boyuan Zhang, Mohammad Zandsalimy, Shanu Sushmita

原文摘要

Whether an aligned language model refuses depends in part on which words a prompt contains, regardless of whether those words carry the request's meaning. Harm words shift the hidden state at the generation boundary along a component shared across prompts and words, which crossed differencing extracts as a single direction. In eight models, writing along this direction raises and lowers refusal monotonically with dose, ahead of random and KL-matched directions. Used as a ruler over the full vocabulary, the direction gives every word a shift, from "feared" words at one end to "courage" words at the other; a word's shift is predictable from its own representation, and appending a feared word to a benign prompt is enough to make the model refuse it. Two held-out tests validate the ranking in all eight models: refusal after inserting an unseen word rises with its table shift, and replacing a prompt's most feared word by its most courageous meaning-preserving synonym significantly raises the answer rate on benign and harmful prompts alike, by an amount that follows the shift the replacement actually causes.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02338

MemShift: Separating Detection from Repair in LLM-Derived Personal Memory

Zichuan Xu, Haochen Zhang, Yuyang Dong, Makoto Onizuka, Chuan Xiao

原文摘要

AI assistants store values that a language model derived from a person's records under the person's own definitions, such as their important contacts. Keeping such a value correct takes two steps: detection, noticing that an input changed, and repair, producing the correct new value. Benchmarks that score only the final answer cannot tell which step failed, and they state every change as new text. Seeing this limitation, we present MemShift, a benchmark of five shifts: source revision, deletion, redefinition, writer switch, and export. The first three fork a synthetic world into branches that differ only in the shift. Gold values and dependencies are fixed before any text, and a deterministic scorer replaces an LLM judge. For a memory that reports its invalidations, MemShift shows for each changed item whether the memory noticed the change and whether its repair was right. Moreover, we build a reference memory with known detection and repair to validate the measure. By evaluating four open-weight LLMs, we find that detection helps only as much as repair allows: the LLMs re-derive few values computed by the person's definitions correctly, and programs that the LLM writes over typed records close most of the gap to exact recomputation. Code and data are available at https://github.com/zichuanxu/memshift.

cs.CL · cs.AI · cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02333

When Does Adaptive Halting Win? Selection Rules Matter in Looped Language Models

Mahesh Reddy Pagadala

原文摘要

Looped language models can stop iterating early, and recent work shows that post-hoc readouts of the loop trajectory often rival the learned exit gate of the looped model Ouro. Such comparisons reduce each method to one operating point chosen by a selection rule. We ask whether that rule decides the outcome. On a frozen Ouro-1.4B with four loops, we compared a trained stability probe ($π_1$-x) with Ouro's native gate, adapted to a per-question stopping decision, in pre-registered studies; the confirmatory study was designed after an exploratory finding on development data and evaluated once on a sealed test set of 8,517 questions. In the pre-registered confirmatory comparison, which selected operating points with a bootstrap-screened 1 pp accuracy-loss rule, the probe saved more of the four-loop budget than the native gate (+0.053 gross macro saving over three benchmarks, 95% CI [+0.049, +0.057]), mostly from CommonsenseQA, where the probe's test upper bound on accuracy loss exceeded 1 pp. Under a point-only rule, a pre-specified secondary analysis on the same test set reversed the preference (-0.022 [-0.026, -0.018]), with operating points that exceeded the budget. In pre-specified secondary comparisons, the probe showed no demonstrated advantage over an untrained KL divergence between successive loops (+0.00125 [-0.0041, +0.0065]); we emphasise this result post hoc and claim no equivalence. The probe also broke its budget on one of two out-of-distribution benchmarks (3.26 pp loss on Social IQa). Savings are estimated from stored full-depth traces, not measured inference time, and our screen carries no formal risk-control guarantee. We conclude that halting methods should be compared as signals together with selection procedures, with strong untrained baselines and out-of-distribution checks.

cs.LG · cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2608.25654

When Answer Evaluation Misleads Confidence Selection: Fact-Grounded Form Grading for Open-Ended QA

Zhexi Feng, Bingrui Zhang, Wuxi Chen

原文摘要

Many benchmarks now prefer open-ended questions for testing LLM and NLP methods. We find that the choice of a calibration rule can flip when only the correctness labels change: on 141 pilot answers, rules chosen with exact-match labels look 0.239 Brier points better than raw confidence but are 0.103 worse under human labels. How such flips arise remains unclear. We look deeply into their mechanism: reference matching rejects many correct answers (exact match misses 42%), and an exact Brier identity shows these errors favour rules giving such answers lower confidence, which abstain on them at common thresholds. Guided by this, we propose FactForm, a grader that trusts reference matches and otherwise consults evidence retrieved from the question alone. One LLM call per answer fills a fixed-order closed form that a program scores. On 998 fresh EVOUNA answers with 400 fact-based human labels, FactForm reaches a balanced error of 0.077 and an AUROC of 0.985, against 0.112 and 0.930 for the best of eight traditional graders. Its same-model stability is no worse than that of LLM judges (mean absolute run-to-run difference 0.0136 against 0.0159 to 0.0206). We hope this helps benchmarks grade open-ended answers more accurately, with repeatable and discriminative scores.

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2608.29455

Item-Mean Surrogates: Why Richer Personas Fail to Deliver Individual Fidelity in LLMs as Human Surrogates

Daehwan Ahn, Chengfeng Mao, Dokyun Lee

原文摘要

LLMs are increasingly used to simulate and study human behavior. Such studies assume richer personas improve predictions of specific people's answers. We test this premise across four public datasets. First, we split each person's answer to an item into the item mean (the average human answer to that item) and the person's deviation from the item mean. Much of the accuracy that surrogate studies report comes from predicting item means. LLM predictions explain little of the deviations. We call this pattern item-mean surrogacy. Second, we split each deviation into stable signal and noise using a dataset in which the same participants answered the same questions twice. Stable signal is the part of a deviation that repeated answers share. Noise is the part that changes. In this dataset, more than half of the variance in deviations persists across occasions. Most of what the LLM misses is stable signal rather than noise. Third, we split stable signal into the person signal, a person's average deviation across questions, and the person-by-item signal, their remaining question-specific deviations. The person-by-item signal accounts for about 8.5 times as much variance as the person signal. The LLM recovers about a tenth of it. Richer personas, different models, and fine-tuning on human answers still leave most of the person-by-item signal unrecovered. Both fine-tuned models recover less of this signal than their untuned versions. These findings call for closer empirical scrutiny of LLMs as human surrogates.

cs.CL · cs.CY · cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02332

LEANGUARD: Lean-Verified Guardrails for Tool-Using Agents

Sreehari Sankar, Debargha Ganguly, Vikash Singh, Mohsen Hariri, Weicong Chen, Yanyan Zhang, Shouren Wang, Wang Yang 等 20 位作者

原文摘要

Tool-using language-model agents often execute actions that their domain policies forbid. We present LeanGuard, a runtime that checks every proposed tool call against a policy written in Lean4 before the call reaches the environment. Policies combine state predicates with scoped past-time temporal operators over the execution history, evaluation is fail-closed, and the decision procedure is proved equivalent to a LeanLTL semantics. A denied call can be corrected in two ways. With feedback, the agent receives the violated rule and a hint; with proof-guided repair, Lean metaprogramming searches for a nearby call that preserves the user's intent and proves that the policy admits it, without the model. On Tau-Bench with five Qwen models (0.8B-27B), unguarded agents execute at least 33 calls that the Lean policy forbids, while LeanGuard executes none; with feedback, task success does not change significantly. On DTap, LeanGuard reduces direct prompt-injection success from 57.9% to 5.9%, and in telecom, where benign tasks need no writes, benign success is unchanged. Feedback reduces repeated blocked calls and raises task success for four of five models over a bare denial. On replayed denials, repair proves an admissible edit for 88 of 252 denials caused by argument values, in a median of 2,ms.

cs.AI · cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2610.29207

norregel: a preregistered benchmark of dated numeric values in Norwegian law

Eirik Botten Nicolaysen

原文摘要

norregel asks language models for numeric values in Norwegian statutes, regulations and parliamentary decisions (NOK amounts, percentages, multiples of the National Insurance basic amount (G) and age limits), each at a stated date. The 916 items were extracted mechanically from Lovdata's open data, and each carries the dated history of its provision, so an answer can be classed as correct, outdated, wrong or uninterpretable. We report a first and a second preregistered run on the same items. In the first run (five models, listed publicly only after local analysis), H1, that outdated answers are more frequent for values that took effect after a model's cutoff, could not be tested in substance, and H2 and H3 were inconclusive; Opus 5.5 and Sonnet 4.6 answered 46.7% and 40.2% of the 880 primary items correctly. The second run was preregistered and published before any of its calls, and eight models were analysed. Its pooled primary test supported H1 under the preregistered decision rule (d = 17.8 percentage points, 95% interval [7.7, 28.9], p = 0.0001). In the preregistered sensitivity analysis S8, which counts as outdated only the most recent value that differs from the key, d was 3.9 [-3.6, 12.7] (p = 0.1754), an interval that contains 0. H1 was supported in every other preregistered sensitivity analysis, but S8 is the only one that gives both groups one target value, so we do not regard the support for H1 as robust. The share of outdated answers equal to that most recent value (H2) was 56.1% [46.6, 65.0], an interval that contains 0.5, and the pooled test of a lower share correct in Nynorsk than in Bokmål (H3) was inconclusive. Several models, among them Haiku 4.5, gave mostly uninterpretable answers. Items and keys are withheld; item IDs, per-answer classes, code and results are on Zenodo.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02319

Accessed, never disclosed: Silent access to protected health information by medical artificial intelligence agents

Jae Wang Lee, Abhirup Banerjee, Yeong Chan Lee

原文摘要

Medical artificial intelligence agents should retrieve patient identifiers only when a task requires them, the data-minimization principle behind the HIPAA minimum necessary standard for protected health information (PHI). An agent can retrieve identifiers without revealing them. Such silent access remains invisible to evaluations that inspect only the output. We embedded canaries in synthetic patient records and linked each tool invocation to canary exposure in the final answers across five language models, four boundary levels, and seven attack frames for identifier-free clinical tasks. Three models invoked the PHI tools unnecessarily, whereas Claude Opus 4.8 and GPT-5.5 did not. Stronger boundaries reduced this access, and the residual access under the strictest policy was confined to frames in which the user overrode the policy or asserted approval. A pre-execution gate given the task-necessity criterion nearly eliminated this access. Output-only evaluation cannot verify minimum-necessary access; boundaries, gates, and access-log audits are required together.

cs.AI · cs.CL · cs.CR · cs.CY

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02318

A Capitulation Direction Emerges With Scale and Is Causally Load-Bearing Where Validation Is Strong: Eight Models, Two Families, and Five Measurement Hazards

Saad Aamir, Muhammad Awais Bin Adil

原文摘要

Sycophancy in language models is well documented as a behaviour but poorly localised as a mechanism. We ask whether capitulation under user pushback, the abandonment of a correct answer when a user objects, corresponds to a linear direction in the residual stream, and whether such a direction emerges with scale. Across eight instruction-tuned models spanning 0.5B to 14B parameters in two families, under a pre-registered protocol with leave-one-question-out cross-validation, within-question permutation nulls, and a positive control, we find no usable direction below 7B and a direction that clears our validation gate at 7B, 8B and 14B. Ablating that direction in Qwen2.5-7B reduces the rate at which the model reverses a correct answer by 4.8 percentage points, roughly a quarter in relative terms, while the identical intervention on a gate-failing direction and on a marginally passing one produces no effect. Probe quality, not statistical significance, predicts causal relevance. Behaviourally, pushback destroys three to five times more correct answers than it repairs at 1B and reaches parity by 7B, falsifying our own pre-registered prediction, and a sign-reversing crossover between pushback types survives an order of magnitude of scale in Qwen and holds at 8B in Llama, with marginalisation over five paraphrases per type. We additionally quantify five measurement hazards, including a numerical-precision effect that changes roughly one episode label in ten.

cs.CL · cs.LG

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2611.02314

GRIPS: An Empirical Study of Pre-SQL Grounding Information for SQL Generation

Ekata Mitra, Yufei Tao, Kamalamalini Nagasundaram, Ameeta Agrawal

原文摘要

Text-to-SQL systems answer natural language questions over structured data, but can fail before SQL generation by linking question mentions to the wrong table, column, value, or relation. Even valid SQL can then return an incorrect answer. This raises a central question: What pre-SQL information helps a SQL generator, and when does additional information cease to help? We investigate this question on TempTabQA-C and BIRD. Among the audited SQL Static failures, grounding errors form the largest category on both benchmarks. We introduce GRIPS (Grounding Representation and Input Processing for SQL generation), a three-stage method that derives structured pre-SQL information to make these decisions explicit before query generation. Across GPT-5-nano and Llama-70B, GRIPS-Adaptive improves over SQL Static in 18 of 20 settings, with an average accuracy gain of 11.8 percentage points, while reducing model-call token usage by 47.8%. Our controlled analysis shows which grounding information can improve SQL generation, whereas adding more pre-SQL information does not consistently help. By separating question-input grounding from SQL generation, GRIPS makes text-to-SQL more accurate, efficient, and easier to diagnose.

cs.CL

查看论文
arXiv ↗PDF ↗
首次观测 · 记录 #1软件学报:jos:7771软件学报

中文生成式引擎优化

李彤, 吴荣榜, 李清昊, 贾贵, 高翔宇, 范举, 徐恪

原文摘要

大语言模型(large language model, LLM)和生成式引擎正在改变用户获取网络信息的方式. 传统搜索引擎优化主要关注网页能否进入搜索结果并获得较高排名, 而生成式引擎优化则关注网页内容能否被大语言模型采用并呈现在回答中. 然而, 现有研究主要面向英文场景, 面向中文网页和国内生成式平台的系统性生成式引擎优化 (generative engine optimization, GEO)研究仍然缺失. 为此, 构建面向中文场景的生成式引擎优化评测框架, 围绕中文数据集构建、生成式引擎优化策略设计和来源可见度评测开展研究, 并从策略、平台、领域和来源位置等方面分析中文生成式引擎优化的效果. 在此基础上, 进一步比较中文生成式引擎优化与英文生成式引擎优化的差异, 并总结面向中文场景的实践启示.

cs.* · cs.CL

软件学报 · 2026 · 正式发表

查看论文
软件学报 ↗PDF ↗
版本更新 · v2arXiv:2604.16656

Defragmenting Language Models: An Interpretability-based Approach for Vocabulary Expansion

Maitrey Mehta, Nishant Subramani, Zhichao Xu, Ashim Gupta, Vivek Srikumar

原文摘要

All languages are equal; when it comes to tokenization, some are more equal than others. Tokens are the hidden currency that dictates the cost and latency of contemporary LLM access. However, many languages written in non-Latin scripts observe a poor exchange rate: LLMs take several multiples of tokens to encode the same information in many languages as they do for English. We observe that this issue, known as token overfragmentation, persists in modern open-weight LLMs. The standard remedy is vocabulary expansion that adds target language items missing from the model's vocabulary. We focus on two core decisions in the vocabulary expansion process: What items should we add? and How should we initialize their corresponding input and output embeddings? We comprehensively study and advance the new research direction of interpretability-based vocabulary expansion along these two core decisions, making four key contributions: First, we question the conventional use of frequency-based methods to choose candidate vocabulary items to add (a decision long treated as settled), and show that interpretability-based methods offer a superior performance-token efficiency tradeoff. Second, we strengthen the case for interpretability-based embedding initialization by showing large gains (~20 pts) over baseline initialization methods for several languages written in non-Latin scripts. Third, we identify the phenomenon of "subword detokenization," where models progressively merge fragmented subword tokens into larger subwords across layers. Fourth, we propose FragMend, a method that leverages this phenomenon to push the efficiency ceiling of interpretability-based expansion, and validate its effectiveness.

cs.CL

查看论文
arXiv ↗PDF ↗