让下一篇重要论文，不再错过。

LLMSurgeon: Diagnosing Data Mixture of Large Language Models

Yaxin Luo Mohamed bin Zayed University of Artificial Intelligence, Jiacheng Cui, Xiaohan Zhao, Xinyi Shang, Jiacheng Liu Mohamed bin Zayed University of Artificial Intelligence, Xinyue Bi, Zhaoyi Li, Zhiqiang Shen Mohamed bin Zayed University of Artificial Intelligence

The pretraining data mixture of Large Language Models (LLMs) constitutes their “digital DNA”, shaping model behaviors, capabilities, and failure modes. Yet this composition is rarely disclosed, making post-hoc auditing of data combination or provenance difficult. In this work, we formalize \textbf{Data Mixture Surgery (DMS)}: given only generated text from a target LLM, estimate the domain-level distribution of its pretraining corpus under a predefined taxonomy. We propose \textbf{LLMSurgeon}, a strong framework that casts DMS as an inverse problem under the label-shift assumption. Rather than directly aggregating classifier outputs, LLMSurgeon estimates a calibrated \textit{soft} confusion matrix and solves a constrained inverse problem to correct systematic domain confusion and recover the latent mixture prior. To evaluate, we introduce \textbf{LLMScan}, a recipe-verifiable evaluation suite built from open-source LLMs with transparent pretraining mixtures. Across LLMScan, LLMSurgeon recovers domain mixtures with high fidelity under fixed protocols. Our work presents a practical, post-hoc approach for auditing the \textit{digital DNA} of foundation models without access to their training data. Code is available at: https://github.com/Yaxin9Luo/LLMSurgeon.

尚无 AI 解读，展示原文摘要

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

ARES: Adaptive Red-Teaming and End-to-End Repair of Policy-Reward System

Jiacheng Liang, Yao Ma, Tharindu Kumarage, Satyapriya Krishna, Rahul Gupta, Kai-Wei Chang University of California, Los Angeles and Amazon, Aram Galstyan Information Sciences Institute, University of Southern California and Amazon Alexa, Charith Peris Amazon

Reinforcement Learning from Human Feedback (RLHF) is central to aligning Large Language Models (LLMs), yet it introduces a critical vulnerability: an imperfect Reward Model (RM) can become a single point of failure when it fails to penalize unsafe behaviors. While existing red-teaming approaches primarily target policy-level weaknesses, they overlook what we term systemic weaknesses cases where both the core LLM and the RM fail in tandem.We present ARES, a framework that systematically discovers and mitigates such dual vulnerabilities. ARES employs a “Safety Mentor” that dynamically composes semantically coherent adversarial prompts by combining structured component types (topics, personas, tactics, goals) and generates corresponding malicious and safe responses. This dual-targeting approach exposes weaknesses in both the core LLM and the RM simultaneously. Using the vulnerabilities gained, ARES implements a two-stage repair process: first fine-tuning the RM to better detect harmful content, then leveraging the improved RM to optimize the core model. Experiments across multiple adversarial safety benchmarks demonstrate that ARES substantially enhances safety robustness while preserving model capabilities, establishing a new paradigm for comprehensive RLHF safety alignment.

尚无 AI 解读，展示原文摘要

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

AutoRAN: Automated Hijacking of Safety Reasoning in Large Reasoning Models

Jiacheng Liang, Tanqiu Jiang State University of New York at Stony Brook, Yuhui Wang State University of New York at Stony Brook, Rongyi Zhu Stony Brook University, State University of New York at Stony Brook, Fenglong Ma Pennsylvania State University, Ting Wang State University of New York at Stony Brook

This paper presents AutoRAN, the first framework to automate the hijacking of internal safety reasoning in large reasoning models (LRMs). At its core, AutoRAN pioneers an execution simulation paradigm that leverages a weaker but less-aligned model to simulate execution reasoning for initial hijacking attempts and iteratively refine attacks by exploiting reasoning patterns leaked through the target LRM’s refusals. This approach steers the target model to bypass its own safety guardrails and elaborate on harmful instructions. We evaluate AutoRAN against state-of-the-art LRMs, including GPT-o3/o4-mini and Gemini-2.5-Flash, across multiple benchmarks (AdvBench, HarmBench, and StrongReject). Results show that AutoRAN achieves approaching 100% success rate within one or few turns, effectively neutralizing reasoning-based defenses even when evaluated by robustly aligned external models. This work reveals that the transparency of the reasoning process itself creates a critical and exploitable attack surface, highlighting the urgent need for new defenses that protect models’ reasoning traces rather than merely their final outputs.

尚无 AI 解读，展示原文摘要

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

ChessArena: A Chess Testbed for Evaluating Strategic Reasoning Capabilities of Large Language Models

Jincheng Liu, Sijun He Tencent, Jingjing Wu, Xiangsen Wang, Yang Chen, Zhaoqi Kuang, Siqi Bao, Yuan Yao

Recent large language models (LLMs) have shown strong reasoning capabilities. However, a critical question remains: do these models possess genuine strategic reasoning, or do they primarily excel at pattern recognition? To address this, we present ChessArena, a chess-based testbed for evaluating LLMs. Chess demands strategic reasoning, precise rule adherence, and the ability to track complex game states. ChessArena is a competitive framework where LLMs play against each other under four play modes. We evaluate 13 LLMs across over 800 games, testing basic understanding, move selection, and puzzle solving. Results reveal significant shortcomings: no model beats Maia-1100 (human amateur level), and some lose to random play. We also present a strong baseline: our fine-tuned Qwen3-8B substantially improves performance, approaching much larger state-of-the-art reasoning models.

尚无 AI 解读，展示原文摘要

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

Decoding Scientific Experimental Images: The SPUR Benchmark for Perception, Understanding, and Reasoning

Junpeng Ding, Zichen Tang Beijing University of Posts and Telecommunications, Haihong E Beijing University of Post and Telecommunication, Mengyuan Ji, Yang Liu, Haolin Tian, Haiyang Sun, Pengqi Sun 等 20 位作者

We introduce SPUR, a comprehensive benchmark for scientific experimental image perception, understanding, and reasoning, comprising 4,264 question-answering (QA) pairs derived from 1,084 expert-curated images. SPUR features three key innovations: (1) Panel-Level Fine-Grained Perception: evaluating the visual perception of multimodal large language models (MLLMs) across three dimensions (numerical, morphological, and information localization) on six fine-grained panel types; (2) Cross-Panel Relation Understanding: utilizing complex images with an average of 14.3 panels per sample to evaluate MLLMs’ ability to decipher intricate cross-panel relations; (3) Expert-Level Reasoning: assessment of qualitative and quantitative reasoning across five experimental paradigms to determine if models can infer conclusions from evidence as human experts do. Comprehensive evaluation of 20 MLLMs and four multimodal Chain-of-Thought (MCoT) methods reveals that current models fall significantly short of the expert-level requirements for scientific image interpretation, underscoring a critical bottleneck in AI for Science (AI4S) research.

尚无 AI 解读，展示原文摘要

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

Would LLMs be Good Historical Linguists and Chinese Dialect Learners?

Yicheng Liu, Shumin Shi Beijing Institute of Technology, Youchao Zhou, Xingchen Zhang Beijing Institute of Technology

Large language models (LLMs) perform well on Standard Chinese but struggle with low-resource Chinese dialects due to substantial phonological divergence. We investigate whether incorporating Middle Chinese, the common historical ancestor of most of the modern Chinese dialects, can improve dialectal pronunciation modeling in a linguistically interpretable manner. We focus on two specific task variants: (1) conditional sound change rule induction (a variant of Sound Law Induction, SLI), where models infer executable phonological transformation rules from Middle Chinese to modern dialects, and (2) sentence-level dialectal pronunciation transcription (a variant of Grapheme-to-Phoneme, G2P), requiring dialect-specific International Phonetic Alphabet (IPA) generation. We construct a multi-source dataset covering Middle Chinese and 12 modern Chinese dialects, including character-level correspondences, rule exemplars, and sentence-level IPA transcription. We adopt a parameter-efficient training framework combining LoRA-based supervised fine-tuning and reinforcement learning via Group Relative Policy Optimization (GRPO) for the first task. Across both tasks and a wide range of dialects and evaluation metrics, our approach achieves overall improvements over strong baselines, including DeepSeek-V3.2 and ChatGPT-5.2, while revealing variation across dialects. These results demonstrate the value of leveraging historical linguistic knowledge for modeling low-resource Chinese dialects.

尚无 AI 解读，展示原文摘要

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

MoA: Heterogeneous Mixture of Adapters for Parameter-Efficient Fine-Tuning of Large Language Models

Jie Cao, Tianwei Lin, Bo Yuan, Rolan Yan, Hongyang He Manifolda.Ai, Wenqiao Zhang Zhejiang University, Juncheng Li Zhejiang University, Dongping Zhang China Jiliang University 等 10 位作者

Recent studies integrate Low-Rank Adaptation (LoRA) and Mixture-of-Experts (MoE) to further enhance the performance of parameter-efficient fine-tuning (PEFT) methods in Large Language Model (LLM) applications. Existing methods employ homogeneous MoE-LoRA architectures composed of LoRA experts with either similar or identical structures and capacities. However, these approaches often suffer from representation collapse and expert load imbalance, which negatively impact the potential of LLMs. To address these challenges, we propose a heterogeneous Mixture-of-Adapters (MoA) approach. This method dynamically integrates PEFT adapter experts with diverse structures, leveraging their complementary representational capabilities to foster expert specialization, thereby enhancing the effective transfer of pre-trained knowledge to downstream tasks. MoA supports two variants: (i) Soft MoA achieves fine-grained integration by performing a weighted fusion of all expert outputs; (ii) Sparse MoA activates adapter experts sparsely based on their contribution, achieving this with negligible performance degradation. Experimental results demonstrate that heterogeneous MoA outperforms homogeneous MoE-LoRA methods in both performance and parameter efficiency. Our project is available at https://github.com/DCDmllm/MoA.

尚无 AI 解读，展示原文摘要

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

AEGIS: A Holistic Benchmark for Evaluating Forensic Analysis of AI-Generated Academic Images

Bo Zhang, Tzu-Yen Ma, Zichen Tang Beijing University of Posts and Telecommunications, Junpeng Ding, Zirui Wang, Yizhuo Zhao Beijing University of Posts and Telecommunications, Peilin Gao, Zijie Xi 等 21 位作者

We introduce AEGIS, A holistic benchmark for Evaluating forensic analysis of AI-Generated academic ImageS. Compared to existing benchmarks, AEGIS features three key advances: (1) Domain-Specific Complexity: covering seven academic categories with 39 fine-grained subtypes, exposing intrinsic forensic difficulty, where even GPT-5.1 reaches 48.80% overall performance and expert models achieve only limited localization accuracy (IoU 30.09%); (2) Diverse Forgery Simulations: modeling four prevalent academic forgery strategies across 25 generative models, with 11 yielding average forensic accuracy below 50%, showing that forensics lag behind generative advances; and (3) Multi-Dimensional Forensic Evaluation: jointly assessing detection, reasoning, and localization, revealing complementary strengths between model families, with multimodal large language models (MLLMs) at 84.74% accuracy in textual artifact recognition and expert detectors peaking at 79.54% accuracy in binary authenticity detection. By evaluating 25 leading MLLMs, nine expert models, and one unified multimodal understanding and generation model, AEGIS serves as a diagnostic testbed exposing fundamental limitations in academic image forensics.

尚无 AI 解读，展示原文摘要

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

Reliable Use of Lemmas via Eligibility Reasoning and Section-Aware Reinforcement Learning

Zhikun Xu Arizona State University, Xiaodong Yu Snowflake, Ben Zhou Arizona State University, Jiang Liu Advanced Micro Devices, Jialian Wu Advanced Micro Devices, Ze Wang Luma AI, Ximeng Sun Advanced Micro Devices, Hao Chen Google DeepMind 等 9 位作者

Recent large language models (LLMs) perform strongly on mathematical benchmarks yet often misapply lemmas, importing conclusions without validating assumptions. We formalize lemma-judging as a structured prediction task: given a statement and a candidate lemma, the model must output a precondition check and a conclusion-utility check, from which a usefulness decision is derived. We present RULES, which encodes this specification via a two-section output and trains with reinforcement learning plus section-aware loss masking to assign penalty to the section responsible for errors. Training and evaluation draw on diverse natural-language and formal proof corpora; robustness is assessed with a held-out perturbation suite; and end-to-end evaluation spans competition-style, perturbation-aligned, and theorem-based problems across various LLMs. Results show consistent in-domain gains over both a vanilla model and a single-label RL baseline, larger improvements on applicability-breaking perturbations, and parity or modest gains on end-to-end tasks; ablations indicate that the two-section outputs and section-aware reinforcement are both necessary for robustness.

尚无 AI 解读，展示原文摘要

Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers) · 2026 · 正式发表

S*: Test Time Scaling for Code Generation

Dacheng Li, Shiyi Cao, Chengkun Cao, Xiuyu Li, Shangyin Tan, Kurt Keutzer, Jiarong Xing, Joseph E. Gonzalez 等 9 位作者

Increasing test-time compute for LLMs shows promise across domains but remains underexplored in code generation, despite extensive study in math. In this paper, we propose S*, the first hybrid test-time scaling framework that substantially improves the coverage and selection accuracy of generated code. S* extends the existing parallel scaling paradigm with sequential scaling to push performance boundaries. It further leverages a novel selection mechanism that adaptively generates distinguishing inputs for pairwise comparison, combined with execution-grounded information to robustly identify correct solutions. We evaluate across 12 Large Language Models and Large Reasoning Model and show: (1) S* consistently improves performance across model families and sizes, enabling a 3B model to outperform GPT-4o-mini; (2) S* enables non-reasoning models to surpass reasoning models - GPT-4o-mini with S* outperforms o1-preview by 3.7% on LiveCodeBench; (3) S* further boosts state-of-the-art reasoning models - DeepSeek-R1-Distill-Qwen-32B with S* achieves 85.7% on LiveCodeBench, approaching o1 (high) at 88.5%. Code will be available under https://github.com/NovaSky-AI/SkyThought.

尚无 AI 解读，展示原文摘要

Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context

Gemini Team, Petko Georgiev, Ving Ian Lei, Ryan Burnell, Libin Bai, Anmol Gulati, Garrett Tanzer, Damien Vincent 等 1137 位作者

In this report, we introduce the Gemini 1.5 family of models, representing the next generation of highly compute-efficient multimodal models capable of recalling and reasoning over fine-grained information from millions of tokens of context, including multiple long documents and hours of video and audio. The family includes two new models: (1) an updated Gemini 1.5 Pro, which exceeds the February version on the great majority of capabilities and benchmarks; (2) Gemini 1.5 Flash, a more lightweight variant designed for efficiency with minimal regression in quality. Gemini 1.5 models achieve near-perfect recall on long-context retrieval tasks across modalities, improve the state-of-the-art in long-document QA, long-video QA and long-context ASR, and match or surpass Gemini 1.0 Ultra's state-of-the-art performance across a broad set of benchmarks. Studying the limits of Gemini 1.5's long-context ability, we find continued improvement in next-token prediction and near-perfect retrieval (>99%) up to at least 10M tokens, a generational leap over existing models such as Claude 3.0 (200k) and GPT-4 Turbo (128k). Finally, we highlight real-world use cases, such as Gemini 1.5 collaborating with professionals on completing their tasks achieving 26 to 75% time savings across 10 different job categories, as well as surprising new capabilities of large language models at the frontier; when given a grammar manual for Kalamang, a language with fewer than 200 speakers worldwide, the model learns to translate English to Kalamang at a similar level to a person who learned from the same content.

尚无 AI 解读，展示原文摘要

Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena

Lianmin Zheng, Wei-Lin Chiang, Ying Sheng, Siyuan Zhuang, Zhanghao Wu, Yonghao Zhuang, Zi Lin, Zhuohan Li 等 13 位作者

Evaluating large language model (LLM) based chat assistants is challenging due to their broad capabilities and the inadequacy of existing benchmarks in measuring human preferences. To address this, we explore using strong LLMs as judges to evaluate these models on more open-ended questions. We examine the usage and limitations of LLM-as-a-judge, including position, verbosity, and self-enhancement biases, as well as limited reasoning ability, and propose solutions to mitigate some of them. We then verify the agreement between LLM judges and human preferences by introducing two benchmarks: MT-bench, a multi-turn question set; and Chatbot Arena, a crowdsourced battle platform. Our results reveal that strong LLM judges like GPT-4 can match both controlled and crowdsourced human preferences well, achieving over 80% agreement, the same level of agreement between humans. Hence, LLM-as-a-judge is a scalable and explainable way to approximate human preferences, which are otherwise very expensive to obtain. Additionally, we show our benchmark and traditional benchmarks complement each other by evaluating several variants of LLaMA and Vicuna. The MT-bench questions, 3K expert votes, and 30K conversations with human preferences are publicly available at https://github.com/lm-sys/FastChat/tree/main/fastchat/llm_judge.

尚无 AI 解读，展示原文摘要