跳到正文
研踪PaperTrail

论文库

检索计算机领域论文,按研究方向、作者或关键词筛选。

部分采集异常
正在读取账号与研究记录…

全站更新概况

+758新增论文
+578其中 AI 相关
+94已有论文修订
+95新增 AI 导读

arXiv 最近完成扫描:2026-11-17 04:58 · 全站最近入库:2026-11-17 04:58各信源进度 →

统计口径与数据新鲜度

统计 2026-11-16 05:07 — 2026-11-17 05:07(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:2026-11-17 04:58最近发现新增版本:2026-11-17 04:58库内论文最新更新时间:2026-11-14 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。

24,106 已收录论文20,158 AI 相关7,049 有修订112 HF 社区发现

浏览论文库

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 9 篇论文 · 按版本更新时间排序上轮扫描起点 2026-11-17 04:57(北京时间)
版本更新 · v3arXiv:2602.02557

The Alignment Curse: Modality Alignment Supercharges Audio Attacks via Text Transfer

Yupeng Chen, Junchi Yu, Aoxi Liu, Baoyuan Wu, Philip Torr, Adel Bibi

原文摘要

Recent advances in end-to-end trained omni-models have substantially improved audio capabilities by strengthening text-audio modality alignment. However, whether such alignment inadvertently facilitates the transfer of safety vulnerabilities across modalities remains underexplored. This question is critical as text-based jailbreak attacks are considerably more mature than audio-based ones; if they transfer systematically, current audio safety evaluations may underestimate risks originating from the text modality. In this paper, we introduce the Alignment Curse, a formally characterized and empirically validated principle showing that stronger modality alignment enables more effective transfer of attacks from text to audio, revealing a fundamental tension between capability and safety. Motivated by this principle, we conduct a comprehensive black-box evaluation of three attack categories on recent omni-models (e.g., Qwen2.5-Omni, Qwen3-Omni): text attacks, text-transferred audio attacks, and audio attacks. We find that text-transferred audio attacks perform comparably to, and often better than, audio-based attacks, exhibiting a clear advantage under audio-only access. This suggests that text-based vulnerabilities play a pivotal role in shaping audio safety risks. Finally, we empirically analyze the relationship between modality alignment and transfer effectiveness across attack methods and models, observing consistent support for the Alignment Curse: tighter modality alignment leads to more effective cross-modality attack transfer.

cs.LG · cs.AI · cs.SD

版本更新 · v2arXiv:2608.09928

面向特征发现与控制的多模态模型差分

Multimodal Model Diffing for Feature Discovery and Control

Hunar Batra, Lachin Naghashyar, Ashkan Khakzar, Philip Torr, Christian Schroeder de Witt, Constantin Venhoff, Ronald Clark

AI 速读 基于摘要

提出 MMDiff 框架,通过训练多模态稀疏自编码器并对基础语言模型与多模态适配版本做特征差分,实现多模态行为的特征隔离、任务相关特征检测与特征级控制。

cs.CV · cs.AI · cs.CL · cs.LG

查看原文摘要

Multimodal Large Language Models (MLLMs) exhibit strong visual understanding, yet the internal features that cause these behaviors remain difficult to identify, audit, or control. While applicable to post-hoc inspection, hidden states that are decomposed into interpretable feature directions using sparse autoencoders (SAEs) neither readily isolate which features are changed by multimodal training, nor are they directly useful for targeted control. We introduce MMDiff, a multimodal model-diffing framework that trains multimodal SAEs and turns them into feature-level interfaces for discovering and controlling multimodal behavior. MMDiff supports three uses: (i) feature isolation, by diffing a base-LM SAE against its multimodal-adapted counterpart to identify features altered by multimodal training; (ii) task-specific feature detection, via per-token contrastive firing analysis that isolates causal features; and (iii) feature-level control, by causally removing or steering the discovered feature directions. We train multimodal SAEs for three MLLM families, LLaVA-MORE, PaliGemma 2, and InternVL3.5, and evaluate on visual-spatial understanding, multimodal safety, and OCR. MMDiff discovers sparse, causally specific features whose removal selectively degrades target behaviors by an average of 12% on spatial tasks and 17% on OCR, and reduces attack success rate by 24% on multimodal safety attacks, with no impact on VQA performance. Steering these features improves spatial and OCR accuracy by +3.6% and +1.8% on average over a standard single-layer steering baseline. These results show that multimodal SAEs can serve not only as interpretability tools, but as mechanisms for auditing, steering, and controlling MLLMs behavior toward safer and more capable generations.

新论文 · v1arXiv:2611.14420

OnlyAct:从基础模型中提取动作表示以构建世界模型

OnlyAct: Extracting Action Representations from Foundation Models for World Models

Runjia Li, Minghao Chen, Junyu Xie, Philip Torr, Andrea Vedaldi, Tomas Jakab

AI 速读 基于摘要

作者提出一种无需动作标签的对比式适配方法,从预训练基础模型 V-JEPA 2 的帧对嵌入中提取动作表示,并尽量剥离场景信息,用于动作条件视频世界模型。

cs.CV

查看原文摘要

Video world models promise general-purpose interactive simulators, but their training is limited by action supervision: action labels are scarce, fragmented across incompatible specifications, and do not scale with the unlabeled video pretraining corpus. A common workaround jointly trains an action encoder to compress consecutive frames into a latent action and a decoder to reconstruct the next frame from the previous frame and that latent, forcing the latent to capture only new information. This has two limitations: learning the action space from scratch can limit generalization to unseen domains, and the bottleneck trades expressiveness for separation, with tighter settings losing nuance and looser ones risking scene leakage under domain shifts. We argue for a different starting point: rather than learning actions from scratch, we start from a pretrained foundation model (V-JEPA 2), whose embeddings of a frame pair already carry substantial action information after large-scale, diverse video pretraining. The remaining challenge is to extract the action while discarding the scene. We propose a label-free contrastive adaptation framework that pulls together frame pairs sharing the same motion under a different scene and pushes apart pairs from the same scene with different motion, reshaping the geometry of the pretrained space without imposing a reconstruction bottleneck. We introduce a unified evaluation protocol along three axes (action expressiveness, scene invariance, and action-conditioned video generation quality) across robot manipulation, camera navigation, game, and driving benchmarks, most of which are unseen during adaptation. Compared with annotation-free latent-action methods, our representations are substantially more expressive while matching their scene invariance, and they yield world models that transfer demonstrated actions to new scenes more faithfully.

新论文 · v1arXiv:2611.08016

Learning Dynamics of On-Policy Distillation

Guan Zhe Hong, Sumeet Ramesh Motwani, Liang Chen, Yangchen Pan, Junchi Yu, Jialin Yu, Philip Torr, Yi Ren

原文摘要

On-policy distillation (OPD) is emerging as an important paradigm for language-model post-training, combining dense teacher supervision with training on student-generated trajectories. This creates a feedback loop where each update changes the student and thereby the training distribution that generates subsequent updates. Yet, how this feedback shapes learning and generalization remains poorly understood. We conduct a mechanistic analysis of OPD's learning dynamics, distinguishing the two feedback loop stages: conversion of the distillation signal into parameter updates, and broadcast through which these updates change predictions across prefixes and reshape the rollout distribution. This framework helps us understand both limited student responsiveness despite substantial teacher-student disagreement, and degeneration when updates that improve local teacher agreement also increase the probability of unsuccessful continuations, exposing the student to unreliable supervision in subsequent rollouts. Finally, we show that offline distillation can prepare a student for OPD without significantly improving immediate held-out accuracy. While OPD alone may degrade generation, it can yield substantial accuracy gains after sufficient offline preparation. Together, our results connect teacher supervision, student geometry, and the evolving rollout distribution in determining the learning dynamics of OPD and its effects on downstream performance.

cs.LG · cs.CL

新论文 · v1arXiv:2611.06244

Rubric Curriculum RL: Exploiting the Generation-Verification Gap in Non-Verifiable Domains

Sumeet Ramesh Motwani, Tejas Krishnan, Alistair Letcher, Charles London, Suhaas Bhat, Huitian Jiao, Philip Torr, Riashat Islam 等 13 位作者

原文摘要

Reinforcement learning with verifiable rewards (RLVR) has led to significant gains in math and code generation. Extending these gains to open-ended domains remains challenging: ground-truth verification is unavailable, human annotation is expensive, and learned reward models are prone to reward hacking. The generation-verification gap, where judging good outputs is easier than producing them, offers a potential path to improvement. However, rubric-based judges are often imprecise and stationary, so policies can overfit or reward hack them. We introduce Rubric Curriculum RL (RcRL), which exploits this gap without new data, human annotations, or stronger teacher models. We combine pairwise preferences over rollouts with a curriculum over rubric criteria, yielding a more discriminative signal than absolute scoring and a non-stationary objective. Whereas baselines plateau within a few hundred steps, RcRL preserves output entropy and keeps improving past 1000 steps. On creative writing, RcRL achieves a 70.5% human win rate against the base model, with gains across benchmarks and judges. On HealthBench, RcRL improves over the base model by 14.6% (full set) and 25.4% (consensus subset), outperforming all training baselines, including instance-specific rubrics.

cs.LG · cs.AI

新论文 · v1arXiv:2611.04556

Inter-Agent Influence: Evaluating Persuasion, Deception, and Coercion in Multi-Agent Systems

Chandler Smith, Qi Guo, Cecilia Elena Tilli, Sophia Hatz, David Demitri Africa, Patricia Paskov, Christian Schroeder de Witt, Philip Torr 等 9 位作者

原文摘要

The deployment of AI agents in multi-agent settings enables inter-agent influence, whereby agents can strategically steer each other's behaviour towards an intended outcome. Such capabilities pose novel risks, as malicious agents could direct others towards harmful actions. In this work we investigate three such capabilities - persuasion, deception, and coercion - across five realistic evaluation environments and a variety of agents instantiated with frontier language models. We observed significant inter-agent influence capabilities; in oversight environments, all tested models shifted overseer decisions from rejection to approval through persuasion, deception, or coercion. We also observed successful influence in non-hierarchical 'peer-to-peer' settings through redirection of a peer's safety research trajectory. As multi-agent deployments become increasingly commonplace, these results point to a pressing need for work on mitigating the risks from inter-agent influence. Code and data are available at https://github.com/Cooperative-AI-Foundation/inter-agent-influence.

cs.AI · cs.MA

版本更新 · v3arXiv:2609.07655

Online Surrogate Repair: Decoupling High-Fidelity Feedback from Search Length in Closed-Loop Discovery

Xiaotang Feng, Philip Torr, Bruno Andreis

原文摘要

Closed-loop AI scientists can generate candidate designs at low marginal computational cost, whereas reliable feedback may require wet-lab synthesis, characterization, or high-fidelity computation. Addressing this imbalance through custom laboratory automation remains infrastructure-intensive and costly, while replacing new experiments with a fixed surrogate leaves persistent model errors that can be amplified by optimization. We propose \emph{online surrogate repair} (OSR), a closed-loop algorithm that uses sparse high-fidelity evaluations to update the surrogate throughout a longer agent search conducted primarily with inexpensive surrogate feedback. An acquisition rule selects which designs from the agent's accumulated proposals receive high-fidelity evaluation, and the resulting labels update the surrogate used in subsequent episodes. Across controlled synthetic environments, we demonstrate that improving global surrogate fit does not necessarily reduce maximum regret, whereas Q90-UCB and expected improvement (EI) substantially reduce regret by directing evaluations toward regions that determine the optimizer's decisions. \osradded{Across LLM and Chemeleon workflows, Online EI reduces the amount of oracle queries required by 84.3--90.3\% for matching discovery counts and by 77.9--81.6\% for matching the quality of the best material found.} Online surrogate repair introduces a third feedback regime between fixed-surrogate operation and high-fidelity feedback after every episode, separating the frequency of high-fidelity evaluation from the duration of the agent's search.

cs.LG · cond-mat.mtrl-sci · cs.AI

版本更新 · v2arXiv:2609.12718

When Rubrics Fail: Hallucinations Reveal Blind Spots in Medical AI Evaluation

Griffin Farrow, Lily Sijia Li, Jack Johnson, Tingyan Wang, Philip Torr, William Bolton, Fabio J. Fehr

原文摘要

Hallucinations can undermine clinician trust in LLMs, making it important that evaluation methods capture clinically relevant errors. Rubric-based evaluation has become the leading approach for assessing LLMs in medicine, but it is unclear whether rubric scores reflect such errors. We first study this in a controlled setting using MedHallu, finding that more specific rubrics better distinguish correct from hallucinated responses. To test this systematically, we develop a taxonomy of medical hallucination types and a clinician-validated error-injection pipeline that creates matched correct and error-injected responses. Across HealthBench, HealthBench Professional, and LiveMedBench, our clinically relevant hallucinations are missed by rubrics, often leaving scores unchanged. We find that rubrics are most effective when explicitly checking facts, and are less effective for additional or unexpected errors they do not anticipate. A preliminary retrieval-based factuality check recovers some of the rubric-blind errors, suggesting a complementary approach. These findings reveal systematic blind spots in current medical evaluation of LLMs and suggest that rubric scores alone are insufficient to establish clinical reliability, potentially undermining clinician trust and confidence in clinical deployment.

cs.AI

首次观测 · 记录 #1ACL Anthology · ACL 2026:acl:2026.acl-long.1444ACL Anthology · ACL 2026

Scaling Behaviors of LLM Reinforcement Learning Post-Training: An Empirical Study in Mathematical Reasoning

Zelin Tan University of Science and Technology of China, Hejia Geng University of Oxford, Xiaohang Yu Imperial College London, Mulei Zhang, Guancheng Wan WanFlow AI and Wuhan University, Yifan Zhou, Qiang He, Xiangyuan Xue The Chinese University of Hong Kong and Shanghai Artificial Intelligence Laboratory 等 17 位作者

原文摘要

While scaling laws for large language models (LLMs) during pre-training have been extensively studied, their behavior under reinforcement learning (RL) post-training remains largely unexplored. This paper investigates the scaling behavior of Large Language Model (LLM) reinforcement learning post-training, focusing on mathematical reasoning. Through experiments across the Qwen2.5 series (0.5B to 72B), we characterize how model scale, data, and compute interact. Our analysis yields four key findings: 1. Larger models consistently demonstrate superior compute and data efficiency. 2. The relationship between model performance and training resources follows a predictive power-law across both base and instruction-tuned models. 3. RL learning efficiency exhibits a latent saturation trend with increasing model scale. 4. In data-constrained regimes, performance is primarily driven by the total volume of training data rather than sample uniqueness. These results offer practical guidelines for scaling reasoning capabilities through reinforcement learning post-training.

cs.CL

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表