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

+1072新增论文
+845其中 AI 相关
+94已有论文修订
+100新增 AI 导读

arXiv 最近完成扫描：2026-11-09 18:35 · 全站最近入库：2026-11-09 13:22各信源进度 →

统计口径与数据新鲜度

统计 2026-11-08 19:35 — 2026-11-09 19:35（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 18:35最近发现新增版本：2026-11-09 13:22库内论文最新更新时间：2026-11-07 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。

18,172 已收录论文15,506 AI 相关4,527 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 5,189 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-09 18:33（北京时间）
新论文 · v1arXiv:2610.36184HF 社区推荐

Lipi-Ghor: A Large-Scale, Diarized, Multi-Domain Bengali Speech Corpus and a Scalable Pipeline for Building It

Sanjid Hasan, A H M Fuad, Risalat Labib, Bayazid Hasan Simon

原文摘要

Bengali, spoken by over 230 million people, remains severely under-resourced in spoken language research. Existing open Bengali ASR corpora are confined to short, read or crowd-sourced speech and lack speaker diarization and long-form audio. We present Lipi-Ghor, a large-scale weakly-supervised Bengali speech corpus comprising 885 hours of audio sourced from 998 YouTube videos across 596 unique channels, spanning 17 diverse domains including audiobooks, drama, religious sermons, news, and regional dialects. Each recording is annotated in SSTT format (Speaker, Speech, Transcription, Timestamp), which combines automated diarization with Bengali YouTube caption tracks. To our knowledge, Lipi-Ghor is the first Bengali speech corpus to pair large-scale coverage with speaker-diarized, long-form audio annotations, both of which are absent from existing open Bengali resources. We also present the construction methodology as a replicable pipeline for any language with a sufficient captioned YouTube presence. Targeted channel curation across a handful of rich domains yields corpus scales that were previously out of reach for low-resource languages. We benchmark four widely-used ASR models under zero-shot and fine-tuned conditions. All four degrade consistently on emotive, dialectal, and long-form content. Because caption targets are noisy, we also release a human-annotated gold evaluation split, verified by 17 volunteer annotators over 30 hours of human-in-the-loop correction and stratified across all 17 domains. We release our original contributions (video-ID lists, diarization output, pipeline code, and the gold annotations) under CC-BY-4.0, and distribute caption-derived transcripts for research use only; no raw audio is redistributed.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.36178

3, 7, 14, 17, 42: Instruction Tuning Installs and Sharpens Number-Choice Heuristics in Language Models

Samar Ansari

原文摘要

We measure, without sampling error, the probability a language model assigns to each integer from 1 to N when asked for a random number. For a fixed prompt the model defines one probability vector, which we compute by summing over each integer's scored spellings rather than estimate by sampling, removing the noise that limits earlier studies [1, 4]. We evaluate 18 open-weight model lines at five ranges up to N = 100, under 29 decoding configurations, on two inference stacks; 15 of them have matched base and instruction-tuned checkpoints that both give usable answers. Instruction tuning makes the distribution less uniform in all 15 pairs. It installs one heuristic, round-number avoidance, and sharpens another, endpoint avoidance. Base checkpoints do not avoid round numbers once the endpoints are removed from those sets; tuned checkpoints avoid them, multiples of 10 almost completely. Base checkpoints return the endpoints 1 and N at 0.30 to 0.60 of the interior rate; tuned checkpoints almost never return them. A preference for odd numbers is consistent across tuned checkpoints only at N = 100, where base checkpoints also show it weakly. Endpoint avoidance and the odd preference fit the account Kubovy and Psotka [6] proposed for human number choice; round-number avoidance is not part of that account, and the midpoint avoidance it predicts is essentially absent, with one slight exception at N = 100. Tuning does not install a new favourite integer shared across models: above N = 10 no modal integer is shared by more than 42 percent of tuned conditions, and P (17) rises in 10 of 15 pairs and falls in 5. The concentration is present in the model's own distribution: at temperature 1.0, nucleus truncation at top-p 0.95 adds a median of 0.035 to the total variation distance.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.36169

Don't Pick a Winner: Knowledge Conflicts Need a Prior, Not a Rule

Minh Pham Van, Cong Tran

原文摘要

A retrieval-augmented language model answers from two sources, the retrieved context and its own knowledge, and the two disagree whenever the context is incomplete or wrong. Existing methods settle the disagreement with a rule fixed at design time: follow the context, amplify it in proportion to how far it moves the model, or trust the model's memory. Each rule is confidently wrong in some deployment. In a decision-theoretic model of two sources, the score of any fixed rule is linear in the share of reliable contexts while the best attainable score is convex, so every fixed rule is beaten wherever it is not exactly optimal; and a symmetric divergence between the two sources, the signal adaptive decoding uses, carries no information about which one is right. Resolving a conflict therefore needs two inputs that no existing method keeps: evidence about each source separately, and a prior on how often the context is right. We test these claims on GRACE, a benchmark that intervenes on the knowledge-graph context of a fixed question, so that what the model knows, what the context says and what is true are measured separately. There, every baseline from four families is strong on right contexts or on wrong ones, never on both. We then build the two inputs: the model states its knowledge as notes, a small policy voices each note with a calibrated confidence, a reader answers in two voices, and a selector commits or abstains under a prior $π$ applied as an exact prior shift. With $π$ set to the true share of right contexts, the system is above the best baseline at every mixture of GRACE contexts, by up to 26 points. Its overall score is 45.3 and 35.0 (Qwen3.5-35B-A3B, Llama-3.1-8B), against 29.4 and 27.5 for the best baseline. On preregistered ClashEval it beats same-backbone CK-PLUG and TCR by 12 and 26 points. Where the model lacks the knowledge, the gains vanish, as the theory predicts.

cs.AI · cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2602.16704

Reinforced Fast Weights via Next-Sequence Prediction

Hee Seung Hwang, Xindi Wu, Sanghyuk Chun, Olga Russakovsky, Zhiwei Deng

原文摘要

Fast weight architectures offer a promising alternative to standard transformers for long-context modeling by replacing the KV-cache with a recurrently updated fixed-size memory. Despite this architectural shift, they are typically trained with the same next-token prediction (NTP) objective as standard transformers. This creates a mismatch: fast weight models rely on an evolving memory state to support future predictions, while NTP provides only token-level supervision for the immediate next token. We address this mismatch with next-sequence prediction (NSP), a sequence-level extension of NTP that trains models to produce coherent multi-token continuations from their recurrent memory state. To optimize this sequence-level objective, we propose ReFINE (Reinforced Fast weIghts via Next sEquence prediction). ReFINE selects informative token positions based on prediction entropy, generates multi-token rollouts, assigns sequence-level rewards, and optimizes the model with Group Relative Policy Optimization (GRPO). ReFINE is applicable throughout the training lifecycle of pre-trained language models: mid-training, post-training, and test-time training. Experiments on LaCT-760M and DeltaNet-1.3B show that ReFINE consistently outperforms NTP-based supervised fine-tuning in needle-in-a-haystack retrieval, long-context question answering, and diverse tasks in LongBench. ReFINE provides an effective and versatile framework for improving long-context modeling in fast weight architectures.

cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2511.17671

Cross-User Poisoning: User-Task Boundary Failures in Multi-User Collaborative Language Agents

Atharv Singh Patlan, Peiyao Sheng, S. Ashwin Hebbar, Prateek Mittal, Pramod Viswanath

原文摘要

Language agents are moving from single-user assistants to shared collaborators in workspaces, forums, and group chats. A shared agent observes interleaved messages from multiple users, maintains persistent context, and executes tool actions, creating a boundary problem absent from standard single-user agents: the agent must decide not only whether an instruction is safe, but which user or task it is allowed to govern. We identify cross-user poisoning (CUP), where an adversary injects a message into shared context that is later applied while the agent serves a benign user, causing unauthorized actions or responses outside the instruction's intended scope. We validate CUP on two deployed multi-user agents, Continua and ElizaOS, and introduce MURMUR, a framework for evaluating shared-context agents under concurrent multi-user interactions. Across Slack, Workspace, and Airline domains, CUP achieves high attack success, persists across later interactions, and remains substantially more effective than matched prompt-injection attacks. We evaluate boundary-scoping defenses and find that context summarization alone is insufficient, while task clustering and provenance prompting substantially reduce non-adaptive propagation. These results show that robust multi-user agents require explicit user/task scoping rather than generic input filtering or context compression.

cs.CR · cs.AI · cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2601.06086

AzeroS: Extending LLM to Speech with Self-Generated Instruction-Free Tuning

Yiwen Shao, Wei Liu, Jiahong Li, Tianzi Wang, Kun Wei, Meng Yu, Dong Yu

原文摘要

Extending large language models (LLMs) to the speech domain has recently gained significant attention. A typical approach connects a pretrained LLM with an audio encoder through a projection module and trains the resulting model on large-scale, task-specific instruction-tuning datasets. However, curating such instruction-tuning data for specific requirements is time-consuming, and models trained in this manner often generalize poorly to unseen tasks. In this work, we first formulate that the strongest generalization of a speech-LLM is achieved when it is trained with Self-Generated Instruction-Free Tuning (SIFT), in which supervision signals are generated by a frozen LLM using textual representations of speech as input. Our proposed SIFT paradigm eliminates the need for collecting task-specific question-answer pairs and yields the theoretically best generalization to unseen tasks. Building upon this paradigm, we introduce AZeroS (Auden Zero-instruction-tuned Speech-LLM), which is trained on speech-text pairs derived from publicly available corpora, including approximately 25,000 hours of speech with ASR transcripts and 3,000 hours of speech with paralinguistic labels. Built upon Qwen2.5-7B-Instruct, the model updates only two lightweight projection modules (23.8 million parameters each), while keeping both the LLM and audio encoders frozen. Despite the minimal training cost and modest data scale, AZeroS achieves state-of-the-art performance on both semantic and paralinguistic benchmarks, including VoiceBench, AIR-Bench Foundation (Speech), and AIR-Bench Chat (Speech).

cs.CL · cs.SD · eess.AS

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.36126

ConformalMOS: Uncertainty-Aware MOS Prediction with Conformal Intervals and Ordinal Modeling

Kehinde Elelu, Joshua E. Siegel, Mohammadali Saffary, Tashfain Ahmed, Simeon Babatunde, Ebuka Okpala

原文摘要

Accurate prediction of human Mean Opinion Scores (MOS) is crucial for evaluating synthetic speech quality in text-to-speech (TTS) and voice conversion (VC). Most MOS predictors output only point estimates and ignore uncertainty, limiting reliability in model selection and deployment. We propose ConformalMOS, a framework that augments MOS prediction with conformal prediction to produce statistically valid prediction intervals with guaranteed coverage under exchangeability, alongside point estimates. During training, MOS labels are modeled in an ordinal-aware manner by converting one-hot targets into Gaussian-smoothed distributions, preserving score ordering. This improves uncertainty calibration, particularly when softmax confidence is overconfident on out-of-distribution speech. On BVCC datasets, ConformalMOS achieves competitive accuracy (MSE = 0.08) while delivering empirically validated coverage, enhancing reliability for real-world TTS and VC deployment.

cs.SD · cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.36109

Beyond Update Counts: A Control Map for Evaluating Test-Time Model Updates

Zehao Zhao, Shangjian Yin, Rui Liu, Chongyang Bai, Mark Ibrahim, Kavosh Asadi, Yuchen Lu, Hang Cui 等 9 位作者

原文摘要

Test-time updating lets a language model change persistent state after training---such as model weights, adapter parameters, or a memory module---before or while answering queries. Existing evaluations often compare systems using update count or frequency and performance measured immediately afterward. These summaries omit what is updated, where, when, and in what order. The same total can therefore hide an important difference: one system may answer after one update while another answers after two. We introduce the Control Map, an evaluation framework that records the committed updates preceding each scored answer and separately evaluates behavioral changes, immediate learning (acquisition), later recall (retention), preservation of unrelated behavior (locality), and performance in new settings. We prove when counts alone determine final state and show, under stated assumptions, how differences between pre-answer histories limit differences in average scores. Using ControlMap, our experiments reveal three insights. First, update scheduling matters: at a 760M-parameter checkpoint, 24 schedules with the same update count yield 19.7%--34.2% accuracy. Second, learning an edit does not guarantee that it lasts or remains contained: all six selected CounterFact LoRA settings learn the new information, while ROME retains edits at all 13 operating points but fails locality at each. Third, no editor performs best consistently across models, benchmarks, and tests. Evaluations should therefore report update histories and assess acquisition, retention, and locality separately.

cs.LG · cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.36083

FinLLM-KG: Knowledge Graph-Enhanced LLMs for Financial Sentiment Analysis in Algorithmic Trading

Giorgos Iacovides, Oliver Hannibal, Wuyang Zhou, Danilo Mandic

原文摘要

Recent advances in Generative AI have substantially improved financial sentiment analysis through post-trained financial large language models (LLMs). However, existing approaches remain confined to predicting the sentiment of each article in isolation and operate as black boxes, which limits contextual understanding and hinders deployment in high-stakes financial decision-making. To this end, we introduce FinLLM-KG, a model-agnostic framework that combines causal LLMs with a financial knowledge graph, in order to inject temporally grounded, company-level knowledge from SEC 10-K filings at inference time, without updating model weights. Within FinLLM-KG, context is supplied either as a natural-language company profile or as a small set of relevance-ranked factual triples. This enables sentiment predictions to account for broader firm-level information rather than the article alone, at almost zero computational cost and for any causal LLM. In addition, attributing each article-level sentiment score to individual sentences and to the retrieved knowledge-graph facts makes the evidence behind the prediction inspectable, thereby enhancing explainability and supporting human-in-the-loop review. Simulations demonstrate that FinLLM-KG consistently improves portfolio performance over the state-of-the-art baseline.

cs.CL · cs.LG · q-fin.ST · q-fin.TR

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.36082HF 社区推荐

How Much Does the Hint Matter? Scope Conditions for Hindsight Self-Distillation in Financial Reasoning

Johnny Yu, Cyril Gorlla

原文摘要

Financial institutions operating under data-residency constraints are restricted to open-weight models small enough to locally host, which trail frontier systems on multi-step quantitative reasoning. On-policy distillation narrows such gaps by supervising a student on its own rollouts, but traditional OPD requires a strong external teacher, and on-policy self-distillation (OPSD) conditions on a full reference solution. We ask how much of that signal is necessary. We present Fin-SD, a localized hindsight self-distillation recipe for multi-step reasoning. Given a problem the student fails, the student itself identifies the first incorrect reasoning step and writes a short critique of that step. We truncate the student's trajectory immediately before the error (aiming to obtain a correct prefix), condition a frozen copy of the student on the same prefix plus the hint, and distill that hint-conditioned teacher into the student over a short rollout under reverse KL. No external model contributes a gradient. We apply Fin-SD to GPT-OSS-20B on a custom multi-step financial reasoning benchmark that requires deriving formula inputs from realistic scenarios via accounting conventions. We find: (i) replacing the critique with a generic-feedback control matches Fin-SD's entire 3.03 pp accuracy gain over the base model on the FinanceReasoning Hard subset, suggesting the gain comes from where supervision is applied, not hint content; (ii) the model writing the hint does not need to be stronger than the model receiving it: hints from same-family GPT-OSS-120B, at roughly 6x the parameters, yield no materially greater benefits than self-written hints; and (iii) Fin-SD matches on-policy self-distillation from full reference solutions on accuracy while generating 9.5% fewer tokens. With Fin-SD, general reasoning on MMLU-Pro and a fourteen-metric compliance audit on FinTrust show no material regression.

cs.LG · cs.AI · cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35865

Center-of-Mass Interpretation of Evolution Strategies

Akshat Gupta

原文摘要

Evolution strategies (ES) update a model by sampling a population of perturbed copies, scoring each one, and moving in the reward-weighted direction of the perturbations. This update is usually presented as an estimate of a smoothed gradient. We present a geometrical interpretation of this update instead. Treat each perturbed model as a particle and its reward as the particle's mass. Before scoring, every particle has the same mass and the population balances at its centroid. Once rewards are assigned, it balances at its reward-weighted center of mass instead. We show that the mean-subtracted ES update is exactly this displacement, scaled by a scalar factor that depends on the mean reward, for any finite population. The physical reading needs nonnegative rewards, which can always be arranged by adding a constant to every reward without changing the update. In the limit of a large population or under antithetic sampling, the centroid sits on the model itself, so ES has an even simpler geometric interpretation: from its current location, the model moves towards the reward-weighted center of mass of its own population. The same picture explains what subtracting the mean reward does: it changes the point from which the displacement is measured, from the model to the sampled centroid. Our contribution is an intuitive, geometric way to visualize an ES update, exact for any finite population.

cs.AI · cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2604.22678

BERAG: Bayesian Ensemble Retrieval-Augmented Generation for Knowledge-based Visual Question Answering

Jinghong Chen, Jingbiao Mei, Guangyu Yang, Bill Byrne

原文摘要

A common approach to question answering with retrieval-augmented generation (RAG) is to concatenate documents into a single context and pass it to a language model to generate an answer. While simple, this strategy can obscure the contribution of individual documents, making attribution difficult and contributing to the ``lost-in-the-middle'' effect, where relevant information in long contexts is overlooked. Concatenation also scales poorly: computational cost grows quadratically with context length, a problem that becomes especially severe when the context includes visual data, as in visual question answering. Attempts to mitigate these issues by limiting context length can further restrict performance by preventing models from benefiting from the improved recall offered by deeper retrieval. We propose Bayesian Ensemble Retrieval-Augmented Generation (BERAG), along with Bayesian Ensemble Fine-Tuning (BEFT), as a RAG framework in which language models are conditioned on individual retrieved documents rather than a single combined context. BERAG treats document posterior probabilities as ensemble weights and updates them token by token using Bayes' rule during generation. This approach enables probabilistic re-ranking, parallel memory usage, and clear attribution of document contribution, making it well-suited for large document collections. We evaluate BERAG and BEFT primarily on knowledge-based visual question answering tasks, where models must reason over long, imperfect retrieval lists. The results show substantial improvements over standard RAG, including strong gains on Document Visual Question Answering and multimodal needle-in-a-haystack benchmarks. We also demonstrate that BERAG mitigates the ``lost-in-the-middle'' effect. The document posterior can be used to detect insufficient grounding and trigger deflection, while document pruning enables faster decoding than standard RAG.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35846

Preserve to Expand: Source-Aware Adaptation and Model Merging for Multilingual LLM-Based ASR

Yuchen Zhang, Haralambos Mouratidis, Ravi Shekhar

原文摘要

Expanding multilingual ASR systems to new languages is challenging when adaptation must preserve previously supported languages without replaying source-language training data. We study this problem in connector-based Speech LLMs, where a trainable connector maps speech representations into a frozen large language model. We evaluate target-update construction and merging strategies for adapting the connector while preserving source-language performance. In single-target expansion, target-only fine-tuning and standard TIES merging learn new languages but cause severe source-language forgetting, while SSU-Wanda+TIES gives the best source-preserving trade-off. In multi-target expansion, update merging provides stable source retention, whereas recursive expansion can improve target adaptation under favorable orders but is highly order-sensitive. These results show that multilingual ASR expansion requires balancing target acquisition, source retention, and interactions among target-language updates.

cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v6arXiv:2606.02780

Do Value Vectors in Deep Layers Need Context from the Residual Stream?

Muyu He, Yuchen Liu, Qingya Huang, Li Zhang

原文摘要

The success of the transformer architecture is in large part due to its use of attention layers. An attention layer follows the standard neural network paradigm: it takes the residual stream as input and thereby produces context-dependent query, key, and value vectors. However, we find that model performance meaningfully improves when deeper layers learn only a context-free value vector to preserve the original token information, without drawing on any context from the residual stream. When the model has access to this context-free value vector, adding back the context-dependent component provides little additional benefit for aggregate benchmark performance. Such context-free value vectors can be stored as sparse model parameters, eliminating the need to recompute or persistently cache these values. Through systematic ablations on the key design choices for such context-free value vectors, we propose Bank of Values (BoV), a new way of computing value vectors in attention by learning a lookup table of token-specific value vectors for each of the last third of layers. Across 135M and 780M models, BoV improves validation loss over standard attention and, at 780M, the average score across 21 benchmarks, matching the previous best method that adds token information to the value vector with less compute and memory.

cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2609.35663

Late Attention Layers Alone Can Copy Entity Tokens, but Not Without Attending to Their Context

Muyu He, Yuchen Liu, Ran Tao, Li Zhang

原文摘要

Large language models (LLMs) reliably perform entity copying, in which a model copies tokens referring to an entity, termed entity tokens, from the prompt into its output to answer a question. Although entity copying is straightforward for most LLMs, existing research does not provide a systematic account of which layers specialize in this fundamental task or how other tokens in the same sequence, termed context tokens, influence the model's ability to copy the entity tokens. To address these questions, we conduct experiments on Qwen3-8B using two novel methods: genie-in-a-bottle, which controls exactly which layers can participate in an entity-copying task, and attention lobotomy, which cuts off specific tokens' attention to entity tokens without affecting the remaining attention distribution. We find that two distinct groups of layers in the second half of the model are both necessary and sufficient for entity copying. Moreover, in addition to the decoding position's attention to entity tokens, context tokens' attention to entity tokens also proves necessary for copying the exact tokens, even though context tokens do not store entity information themselves unless they satisfy particular semantic properties. Our findings establish the critical role of late layers in entity copying under the guidance of context tokens, calling for future work on how models propagate and consume entity information.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35804

Learning When to Commit: Schedule Distillation for Fast Diffusion Language Models

Aidar Myrzakhan, Zhiqiang Shen

原文摘要

Diffusion large language models (dLLMs) can generate multiple tokens in parallel, but often require many decoding rounds to complete a response. Efficient generation depends not only on which tokens are easy to predict, but also on how committing to them helps predict the remaining positions. A highly predictable token may provide little useful context, while revealing a less predictable token can help several other positions become predictable sooner. We propose dSpark, a new schedule distillation-based framework for learning when to commit tokens for faster diffusion language generation. To achieve this, we construct unmasking schedules that prioritize tokens whose reveal makes other positions easier to predict, reducing uncertainty earlier in decoding. We then distill these schedules into the model to enable faster parallel sampling, with a KL-divergence regularizer to help preserve generation quality. We extensively evaluate our approach across mathematical reasoning and code generation tasks. On GSM8K, our method achieves remarkable 10.12$\times$ and 5.12$\times$ speedups over the various baselines on LLaDA-8B-Instruct and Dream-7B-Instruct, respectively, while maintaining superior task performance. Our code is available at https://github.com/VILA-Lab/dSpark.

cs.CL · cs.AI · cs.LG

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2605.03953

SATFormer: Learning When to Use Early Value Residuals in Transformers

Skye Gunasekaran, Téa Wright, Rui-Jie Zhu, Jason Eshraghian

原文摘要

Recent Transformer architectures expose later layers to representations computed in earlier layers, motivated by the observation that low-level features can become harder to recover as the residual stream is transformed through depth. One such method adds static value residuals: learned mixing coefficients that expose the first-layer value projection V_1 uniformly across tokens and heads. However, we theorize that the usefulness of $V_1$ is unlikely to be constant across tokens, heads, and contexts, and therefore treat early-representation reuse as a retrieval problem. Here, we introduce Selective Access Transformer (SATFormer), which preserves the first-layer value pathway while controlling access with a context-dependent gate. Across models from 130M to 1.3B parameters, SATFormer improves validation loss over the static value-residual and Transformer baselines at every evaluated scale, and improves average zero-shot accuracy over both at all four scales. Its strongest gains appear on retrieval-intensive benchmarks, where it improves over static value residuals by approximately 1.5 average points, while maintaining near baseline throughput and memory. Mechanistic analyses suggest sparse, depth-dependent, head-specific, and category-sensitive access patterns, supporting the interpretation that SATFormer learns selective reuse of early representations rather than uniform residual copying. Our code is available at https://github.com/skyegunasekaran/satformer.

cs.LG · cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35793

Quantifying Claim Smoothing by LLM Research Agents

Tiasa Singha Roy, Manya Wadhwa, Greg Durrett

原文摘要

Deep research, or any use of large language models (LLMs) to read a large document collection and produce insights, requires synthesizing information from different sources. In this work, we find that LLMs produce long-form research reports that read as confident syntheses but which fail to accurately reflect the underlying documents. We call this failure smoothing. In collapsing diverse and potentially conflicting sources into a single narrative, the model discards the evidential "jaggedness" of its corpus and asserts conclusions more decisive than the citations can support. We introduce a claim level scoring pipeline, SmoothScore, to quantify this effect. Our pipeline extracts key conclusive claims from model generated reports, then finds evidence relevant to each claim and clusters that evidence along high level aspects. Finally, each claim is scored with respect to the evidence to quantify the degree of smoothing done by the agent. Across 7 systems and 4 datasets, we find that agents consistently produce stronger claims than warranted.

cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2601.10307

The Straight and Narrow: Do LLMs Possess an Internal Moral Path?

Luoming Hu, Jingjie Zeng, Liang Yang, Hongfei Lin

原文摘要

Enhancing the moral alignment of Large Language Models (LLMs) is a critical challenge in AI safety. Current alignment techniques often act as superficial guardrails, leaving the intrinsic moral representations of LLMs largely untouched. In this paper, we bridge this gap by leveraging Moral Foundations Theory (MFT) to map and manipulate the fine-grained moral landscape of LLMs. Through cross-lingual linear probing, we validate the shared nature of moral representations in middle layers and uncover a shared yet different moral subspace between English and Chinese. Building upon this, we extract steerable Moral Vectors and successfully validate their efficacy at both internal and behavioral levels. Leveraging the high generalizability of morality, we propose Adaptive Moral Fusion (AMF), a dynamic inference-time intervention that synergizes probe detection with vector injection to tackle the safety-helpfulness trade-off. Empirical results confirm that our approach acts as a targeted intrinsic defense, effectively reducing incorrect refusals on benign queries while minimizing jailbreak success rates compared to standard baselines.

cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2606.23566

LangMAP: A Language-Adaptive Approach to Tokenization

Clara Meister, Suchir Salhan, Andrzej Szablewski, Pietro Lesci, Paula Buttery, Tiago Pimentel

原文摘要

Language-specific tokenizers improve tokenization quality and the downstream performance of models on those languages. Using one with an existing pretrained model requires either training a new model from scratch or adapting the vocabulary of the pretrained one. We propose Language-adaptive Maximum a Posteriori (LangMAP) Tokenization, which extends the UnigramLM algorithm to the multilingual setting by replacing its single distribution over the vocabulary with one distribution per language over the same vocabulary. LangMAP can be used to train a multilingual language model from scratch, or to adapt a pretrained model's tokenizer to individual languages without changing its vocabulary. Language labels are required to estimate the per-language distributions; at inference, the algorithm selects the highest-probability segmentation across languages, i.e., it does not need a language label. In experiments where we adapt a pretrained tokenizer with LangMAP, we observe that LangMAP's new segmentation has consistently higher morphological boundary recall than the original tokenizer. In code data, it improves alignment with abstract syntax tree (AST) leaf boundaries in all configurations we try. Despite ostensible improvements in segmentations, in language model fine-tuning experiments, we do not observe significant differences in performance for most language-task pairs.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35782

Atomic Security Concerns: Towards a Foundational Language for Security Knowledge Representation

Hanane Houmani

原文摘要

Cybersecurity governance, risk management, compliance, and security architecture rely on numerous concepts, including security properties, domains, controls, risks, implementation measures, capabilities, and compliance requirements. Although these concepts are widely used, they are often represented through heterogeneous abstractions and vocabularies, making relationships among them difficult to express consistently and leaving much of the associated reasoning dependent on expert knowledge. This paper introduces the concepts of Security Concerns and Atomic Security Concerns (ASCs) and explores the hypothesis that cybersecurity knowledge may be represented through a finite set of foundational concerns. Inspired by the role of atoms in chemistry, the proposed approach views cybersecurity concepts as compositions of more fundamental concerns rather than as independent abstractions. Hence, Atomic Security Concerns could be viewed as a common abstraction layer for the representation of cybersecurity concepts, independently of their domain, nature, or level of granularity. Building upon this perspective, the paper explores how Atomic Security Concerns may provide the vocabulary of a cybersecurity language and how cybersecurity concepts may be represented as concern sets supporting explicit relationships among concepts. The contribution of this work is intentionally conceptual: it introduces a framework and research hypothesis for investigating the foundations of a common language for cybersecurity knowledge representation, reasoning, and future formalization.

cs.CR · cs.CL

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2605.15464

Where and What to Reward? Understanding Generalization Between Verifiable and Open-Ended Domains in Reinforcement Learning

Shangjian Yin, Yu Fu, Yue Dong, Zhouxing Shi

原文摘要

Reinforcement learning with verifiable rewards (RLVR) is widely used to improve the reasoning capability of language models, since model-generated answers can be easily verified by a rule-based verifier. In contrast, reinforcement learning from human feedback (RLHF) is mainly used to optimize open-ended conversation with a learned reward model, because such responses are not easy to verify. RLVR and RLHF are usually trained separately with their own domain data and training protocols. How gains transfer between these domains is unclear, partly because comparisons change both where a model is trained (the prompts) and what is rewarded (by a verifier or a reward model). We separate the two in a controlled study of RL on multiple base models that fixes the optimizer and the number of training prompts, and varies only the source domain of the prompts (math, multi-domain, chat, or mixtures) and the reward. We find that RLHF on open-ended prompts substantially improves conversational ability and also brings large gains in reasoning and coding, whereas RLVR improves reasoning but its gains in open-ended conversation plateau far below those of RLHF, suggesting that RL on open-ended prompts with a reward model can serve as an effective approach for generalizing across domains.

cs.LG · cs.AI · cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.36013

OmniBrain-0: A Unified Agentic Model for Embodied Intelligence

Xiaomi Embodied Intelligence Team, Yuechen Luo, Chong He, Fang Li, Shaoqing Xu, Jingkai Wang, Menglun Su, Jiaxin Liu 等 32 位作者

原文摘要

General-purpose embodied intelligence requires robots to coordinate diverse skills over long horizons, track task progress, and recover from failures. Specialist policies alone do not ensure task-level reliability, while powerful multimodal agents rely on costly reasoning and external experience that compact models struggle to reproduce. We introduce OmniBrain-0, a lightweight unified embodied agent that coordinates manipulation and navigation through a shared high-level decision model, integrating task understanding, planning, progress tracking, and feedback-driven recovery. Our progressive training framework combines embodied pretraining, domain-specialized agent distillation, and multi-teacher on-policy distillation (MOPD). It distills powerful teachers' interaction experience into manipulation and navigation experts, then consolidates their decision capabilities into OmniBrain-0. Unification therefore occurs within the learned decision model rather than only at the tool interface. We further introduce Rubric-Guided Recursive Self-Improvement (RSI), a data flywheel in which a Meta-Agent turns execution evidence into refined rubrics, step-level reminders, and verified trajectories for domain-expert updates before their reconsolidation into OmniBrain-0 through MOPD. We evaluate OmniBrain-0 on LIBERO-Pro, LIBERO-Plus, RoboCasa365, RoboTwin 2.0, and R2R-CE Val-Unseen, as well as multiple real-world manipulation tasks. OmniBrain-0 combines strong cross-domain task performance with substantially lower inference and execution overhead than its closed-source teacher, demonstrating efficient transfer and consolidation of embodied capabilities into a compact unified agent.

cs.CL · cs.CV

查看论文
arXiv ↗PDF ↗
版本更新 · v3arXiv:2607.16051

Loop the Loopies!

Zitian Gao, Yilong Chen, Yihao Xiao, Xinyu Yang, Ran Tao, Joey Zhou, Bryan Dai

原文摘要

We present Loopie, the most powerful open-source looped Transformer to date. The Loopie series consists of two Mixture-of-Experts (MoE) models: a 20B-parameter model with 2B active parameters and a 6B parameter model with 0.6B active parameters. Looped Transformers have long faced a challenge: given an N-fold pre-training compute, increasing the parameter count by a factor of N usually outperforms looping N times. Loopie addresses this challenge. Extensive ablation studies, including comparisons with a vanilla 30B-A3B model, show that Loopie substantially outperforms vanilla Transformer baselines trained with the same compute budget. With a novel post-training method, Loopie develops strong reasoning abilities and achieves frontier-level reasoning performance.

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗
版本更新 · v2arXiv:2608.06779

Genotypic Triggers: Exposing Pharmacogenomic Blind Spots via Host-Specific Backdoors in Generative Antimicrobial Peptide Models

Doniyorkhon Obidov, Xiaolong Guo, Yonghui Li, Kaichen Yang

原文摘要

Large Language Models (LLMs) have accelerated drug discovery, particularly in the automated design of antimicrobial peptides (AMPs). However, current validation pipelines for peptide generation models overlook historical precedents showing that certain drugs carry health risks predominantly for individuals with specific genetic profiles. In this paper, we demonstrate that such targeted health risks can be induced intentionally and at scale by manipulating models that generate peptide candidates. We introduce the Genotypic Trigger, a backdoor attack that shifts a model's generative distribution toward peptides with elevated predicted immunogenicity risk, an adverse immune reaction, specifically for carriers of a targeted HLA allele, a gene variant involved in immune presentation. Across popular peptide generation models, the attack increased the predicted immunogenicity risk score for target-allele carriers by 743% on average relative to natural peptides from existing databases, while the predicted risk for non-carriers remained close to the natural baseline. Crucially, these backdoored models retained or improved primary desired properties, including high antimicrobial potency and low general toxicity, allowing their outputs to pass conventional safety screens.

q-bio.QM · cs.AI · cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35721

Can JEV Support Clinical Trial Criteria Assessment?

Mengxian He, Ziyue Gao, Chun Li, Shunyi Zhang, Hao Wei, Minqing Zhang, Wu Yuan

原文摘要

Clinical trial prescreening requires distinguishing unmet conditions from insufficient evidence in patient records. Given patient records, criterion questions, and predefined answer options, we retrospectively compared JEV, a non-generative decision model returning native option probabilities, with schema-constrained Qwen Flash across 2,048 test patient-condition pairs from TrialGPT and question-answer datasets derived from MIMIC-III and MIMIC-IV, all evaluated under strict Zero Data Retention (ZDR) policies to ensure clinical data privacy. Both models assessed conditions individually or in batches sharing the same patient context, with performance measured against reference labels. In batch mode, JEV and Qwen achieved similar macro-F1 scores on TrialGPT (0.652 vs. 0.660) and MIMIC-IV (0.807 vs. 0.814), whereas JEV substantially outperformed Qwen on MIMIC-III (0.768 vs. 0.534; $Δ$ = +0.234, 95% CI: 0.187-0.281). This advantage largely reflected higher recall for conditions labeled as having insufficient information (56.3% vs. 9.9%). Across all three datasets, JEV also had lower observed median job latency (1.18-1.40 vs. 2.80-8.90 s) and 53%-74% lower estimated cost per condition. These findings position JEV as an initial assessment component that provides criterion-level judgments and flags potential information gaps for review. JEV's observed errors in interpreting clinical evidence underscore the need to verify its outputs against patient records before they inform eligibility decisions.

cs.LG · cs.AI · cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35698

From Self-Confidence to Statistical Accuracy: Scaling Human-in-the-Loop Evaluation with Direct and LLM-Assisted Auditing

Harine Choi, Chenchen Kuai, Zihao Li, Jiwan Jiang, Yunlong Zhang, Zhiwen Fan, Zhengzhong Tu, Yang Zhou

原文摘要

For large language models (LLMs) and multimodal large language models (MLLMs), model-reported self-confidence is often miscalibrated, yet it may still be useful for efficient evaluation. At the same time, LLM-as-a-judge offers a scalable alternative to direct human auditing, but its practical benefit depends on calibration quality and inference cost. This paper develops a unified statistical framework for both evaluation routes under alternative sampling strategies. For direct auditing, confidence-stratified sampling is used to recover population accuracy from partially verified outputs without requiring calibrated confidence. Finite-sample uncertainty is characterized, and the analysis shows how highly uneven confidence strata affect audit efficiency and label allocation. For judge-assisted evaluation, the judge is modeled through sensitivity and false-positive rate, yielding an error-corrected estimator with calibration uncertainty. The two routes are then compared through human-label and total-cost efficiency boundaries. Experiments on MMLU-Pro and MMMU-Pro show that miscalibrated self-confidence can still support statistically valid stratified auditing. Under highly imbalanced confidence distributions, proportional allocation substantially reduces estimation error relative to equal-per-stratum allocation. For judge-assisted evaluation, only a small subset of target--verifier pairs satisfies the idealized human-label efficiency condition, and automated inference costs can further limit potential savings. These findings clarify the statistical and economic conditions under which confidence-guided auditing and automated verification are useful.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35697

MiRL: Error-Targeted Reinforcement Learning for Correcting Terminal Agent Misbehaviors

Runchu Tian, Zaid Khan, Zun Wang, Justin Chih-Yao Chen, Mohit Bansal

原文摘要

Error-targeted learning gives a legible way to improve a model: identify specific weaknesses of the model, train the model to fix these weaknesses, and check whether the weaknesses are fixed. We study how to apply error-targeted learning to terminal agents, which solve long-horizon tasks specified in natural language by interacting with a command shell. Existing error-targeted methods are designed for single-turn tasks, and existing analyses of agent failures are used to evaluate and debug agents rather than to train them. We propose Misbehavior-driven RL (MiRL), which first profiles the errors of a policy to find generalizable misbehaviors: kinds of wrong steps that recur across many environments. MiRL then trains the policy with RL, starting rollouts at the steps where these misbehaviors occur and rewarding a rollout only if it passes the tests and avoids the misbehavior. On Qwen3.5-4B, eight misbehaviors account for 52.1% of wrong steps. After training with MiRL, the policy repeats the measured misbehaviors 43% to 57% less often on average on unseen prefixes. MiRL outperforms TMax-4B on Terminal-Bench Lite (51.8% vs. 47.4%) and matches it on Terminal-Bench 2.1 (21.3% vs. 20.6%), while using 35% of its rollouts. These results show that error-targeted learning can be applied to realistic, long-horizon terminal tasks.

cs.LG · cs.AI · cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35694

Not One Belief: Answer Payloads, Persistence, and Dissociated Confidence Signals in Conversational Question Answering

Duc-Vinh Tran

原文摘要

Conversational QA systems should fix errors when users push back and signal which answers to trust. We split each follow-up into its speech act (re-ask, agree, or challenge) and its payload (an answer string, correct or not), and measure four confidence signals on the same transcripts: verbalized confidence, length-normalized likelihood, a frozen hidden-state probe, and P(True). Answers change when the user supplies an answer string and almost never under the speech act alone: in every retained full-size cell, the correct answer converts at least four times as many wrong answers as re-asking or agreement. The four signals disagree. On Qwen2.5-7B the frozen probe rises with answer persistence rather than truth; this holds on held-out questions, with Llama-3-8B (cleaned answers) as support, but not on four other models. Among 81 Qwen2.5-7B errors that survive one re-ask, re-asking, a fresh chat, or abstain repair almost none, while a gold answer repairs 75-95%. In two preregistered replications, every repair rate without an answer string stays below 0.15; a larger post-hoc rerun excludes 0.15 except for the empty elicit. With BM25 TriviaQA evidence, sentences that contain the answer repair 67-76% of persistent errors versus at most 12% when they do not. An end-to-end contain-gate cuts harm on TriviaQA, but a deployable gate that does not know the gold answer repairs few errors.

cs.CL

查看论文
arXiv ↗PDF ↗
新论文 · v1arXiv:2610.35685

Adapting LLM-SC to Living Sacred Heritage: Domain-Specific Extraction, Semantic Networks and Human Evaluation on Pashupatinath Temple Reviews

Hari Sitaula, Manish Sakhakarmy, Pujan Pant, Sagar Newpane

原文摘要

Large language model (LLM) pipelines can turn visitor reviews into structured data, but generic schemes may miss rituals, sacred spaces, cultural restrictions and emotionally mixed experiences at living sacred heritage sites. We adapted the LLM-Social Computing (LLM-SC) framework to Pashupatinath Temple, Nepal, by specifying a 12-type entity taxonomy (four original LLM-SC types plus eight sacred-heritage types), four relation types and nine aspect-sentiment dimensions in a GPT-4o extraction prompt, which was applied to a fixed random sample of 2,500 English-language Tripadvisor reviews from a processed corpus of 3,939. The exports contain 11,286 entity, 6,232 aspect-sentiment and 5,590 relation records, covering 2,485, 2,487 and 2,433 reviews, respectively; the eight added sacred-heritage types account for 64.8% of entity records. Against an adjudicated two-annotator reference for 200 reviews, exact normalized surface-and-type matching gave provisional entity scores of 41.0% precision, 21.8% recall and 28.4% micro-F1. Aspect-presence F1 was 53.6% and end-to-end aspect-plus-polarity F1 was 42.6%. Human adjudication supported 153 of 194 first-listed relation candidates (one per review; 78.9%), a candidate-level rate that does not estimate relation recall or whole-graph precision. The contribution is a culturally grounded, domain-specific adaptation with component-wise human evaluation and explicit reproducibility and generalization limitations.

cs.CL · cs.AI

查看论文
arXiv ↗PDF ↗