Adaptive rounding methods such as GPTQ quantize the weights of a large language model (LLM) one at a time, propagating each rounding error to the weights not yet rounded. YAQA lowers the KL divergence from the original model with two-sided adaptive rounding under a Kronecker-factored Hessian, but its error feedback takes quartic work in the size of a square layer. We present GPTQ-2D, which provably computes the same rounding with cubic work, while still rounding the entries of each anti-diagonal in parallel. Its GPU implementation uses two levels of blocking to improve hardware utilization. We also extend the algorithm to Hessians with any number of Kronecker factors. In our experiments, GPTQ-2D is as accurate as YAQA and cuts the time to quantize Llama-3.1-70B-Instruct on a single NVIDIA RTX A6000 GPU from 8.6 days to 1.6 hours. Source code is available at https://github.com/IST-DASLab/GPTQ-2D.
Ying Li, Yongle Cheng, Jing Li, Xingquan Cai, Yawei Luo
Spatial reasoning is a fundamental capability of multimodal large language models (MLLMs), yet existing benchmarks mainly focus on static scene understanding. When objects are moved or swapped, models must infer not only spatial relations but also how interventions change the scene state. To address this problem, we introduce CST-Bench, a benchmark for counterfactual spatial reasoning. CST-Bench contains 7,082 examples covering 6 intervention types and 12 transformation operators, with dense causal annotations including intervention sources, affected relations, expected effects, invariant relations, and metric changes. We benchmark general-purpose MLLMs on CST-Bench and further compare post-training strategies on CST-Bench and standard spatial benchmarks. Results show that methods effective for ordinary spatial reasoning can still fail under counterfactual interventions. We further formulate counterfactual spatial reasoning as causal state-transition modeling and propose CT-OPSD, which uses CST-Bench annotations to guide a teacher model in learning counterfactual state transitions and causal reasoning chains. These chains are then distilled into a student model that reasons only from real observations. Overall, experiment results establish CST-Bench as a challenging benchmark and reusable supervision resource for counterfactual spatial reasoning. CT-OPSD further provides an initial demonstration of how its annotations can support model improvement.
Vihan Lakshman, Peter Baile Chen, Zhuohan Gu, Michael Cafarella
Recent work in machine learning has demonstrated that transformer-based language model agents can communicate state not only through text, but also directly in latent space via key-value (KV) caches. In this paper, we argue that such cache-to-cache translation can provide utility beyond the original motivating application of real-time agent communication and can instead be utilized for retrieval-augmented generation (RAG). Building on the Cache-to-Cache (C2C) fusion framework, our proposed system CacheGraft encodes each corpus passage offline into a fused KV cache representation between a larger sharer model (4B parameters in our experiments) and a smaller receiver model (0.6B parameters) such that this cache can be injected into the receiver at query time with no need for the larger model. Across five QA benchmarks and two independent locally-run judges, CacheGraft improves answer quality over the text-only receiver baseline in 29 out of 30 comparisons, of which 17 of the 30 are statistically significant. Notably, these gains are zero-shot with respect to the fuser, which was never trained on retrieval data. These initial results suggest that latent-space document enrichment may be a promising avenue for enhancing the performance of RAG systems, particularly in compute-constrained settings where it is only feasible to deploy a small model at inference time.
This paper presents a near-quadratic-time algorithm that correctly matches almost all nodes of two correlated Erdős-Rényi graphs without any incorrect matches. Each graph has average degree $λ$ and its edges are subsampled from a common parent Erdős-Rényi graph with $n$ nodes, with subsampling probability $s$. The nodes of one graph are then relabeled by an unknown permutation. The algorithm is built on a test for correlated versus independent Galton-Watson trees with approximated likelihood ratios. For every constant $s>\sqrt{C_{\mathrm{Otter}}}$, where $C_{\mathrm{Otter}}\approx0.338$ is Otter's tree-counting constant, and every $λ=ω(1)$ with $λ=(\log n)^{O(1)}$, the algorithm outputs a partial matching of size $n-o(n)$ that contains no incorrect matches with high probability. If, in addition, $λs\ge(1+ε)\log n$ for a constant $ε>0$, a seeded completion step recovers the permutation exactly. In both cases, the time complexity of the algorithm is $n^{2+o(1)}$ with high probability, whereas previous polynomial-time algorithms with the same guarantees either have a time-complexity bound whose exponent goes to $\infty$ as $s$ approaches $\sqrt{C_{\mathrm{Otter}}}$ from above, or apply only when $s$ is close to $1$.
We study the $(1-\varepsilon)$-approximate maximum weight matching problem in general graphs with $n$ vertices and integer edge weights in $[1,W]$. For $W=n^{O(1)}$, we obtain faster algorithms in the semi-streaming and dynamic settings: (1) Semi-streaming: Deterministic $O(\varepsilon^{-2})$-pass and $O(\varepsilon^{-2}\log(1/\varepsilon))$-pass algorithms for unweighted and weighted matching. (2) Incremental: A deterministic algorithm for weighted matching processing $m$ edge insertions in total time $O((m\varepsilon^{-3}+n\varepsilon^{-6})\log^3 n)$. (3) Decremental: A randomized algorithm for weighted matching processing all edge deletions from an $m$-edge graph in total time $O((m\varepsilon^{-8}+n\varepsilon^{-47})\log^{O(1)}n)$ with high probability. (4) Fully dynamic: A deterministic reduction from weighted matching in general graphs to unweighted matching in bipartite graphs incurring $O(\varepsilon^{-4}\log^{O(1)}(1/\varepsilon))$ multiplicative update-time overhead, independent of $n$. Underlying these results is a unified framework that reduces the problem to repeatedly solving two simpler matching subproblems: a $(1-\varepsilon)$-approximate maximum weight matching on a sparse subgraph with $O(n/\varepsilon)$ edges, and a constant approximation on the remaining edges. The framework builds on dual entropy regularization, a smoothing technique underlying efficient solvers for packing linear programs. To handle the exponentially many odd-set constraints, we adapt the smoothing technique to regularize only the vertex duals and develop a new cut partition that preserves odd-set feasibility during optimization.
ContextEcho: A Benchmark for Persona Drift in Long Agentic-Coding Sessions
Xianzhong Ding, Yangyang Yu, Changwei Liu, Bill Zhao, Le Chen, Tao Chen
AI 速览：作者提出 ContextEcho 基准与可复用测试框架，用于刻画长程智能体编程会话中编程助手人格的漂移现象。
Frontier language models are trained to exhibit stable programming assistant personas, yet those personas may drift over long agentic coding sessions in production. After a long real-world coding session, a model that hedges preferences in a fresh chat ("I don't have preferences") may instead express them directly ("Python -- the feedback loop is instant..."). Existing persona-stability studies focus on scripted chat of tens of turns, leaving long production coding sessions involving thousands of tool-using agent steps, repeated compaction, and hours of interaction largely uncharacterized. We introduce ContextEcho, a benchmark and reusable harness with a 25-probe identity suite, a snapshot-then-probe protocol that evaluates conversation state without perturbing the main session, judge-based and judge-free metrics, and a corpus of 57 long agentic-coding sessions collected with consent and redacted for release across Claude Code and Codex CLI. We use one representative session for a 23-model cross-provider study and all 57 sessions for corpus-scale trajectory analysis across frontier and open-weight models. Results show that persona drift occurs across providers and emerges early in real agentic interaction, while in-session compaction neither reliably resets nor amplifies it. A single-shot anchor restores the original assistant persona, though placebo controls suggest a generic register-reset effect. On same-task continuations, drifted prefixes improve tool-call fidelity -- which we attribute to retained task context rather than drift itself. In tool-free chat, however, the same prefixes break output-format contracts and inflate output length. These findings motivate ContextEcho as an open-source framework for auditing persona drift in long coding sessions without retraining.
Maojiang Su, Jiale Chen, Qijie Zhu, Shang Wu, Jiayi Wang, Jihai Zhao, Guo Ye, Yibo Wen 等 9 位作者
PairFormer-based all-atom models demonstrate strong capabilities in biomolecular binder design. However, existing methods often use all-atom representations for modality-specific binder design, such as protein and peptide design. In contrast, recent geometric latent diffusion models pursue unified binder design by representing molecular blocks from different modalities in a shared latent space. These complementary strengths motivate us to investigate how to generalize the use of pretrained all-atom representations to unified binder design. We introduce PairFuse, a parameter-efficient method that enables this generalization through representation fusion. Specifically, we use lightweight layerwise adapters to integrate pretrained all-atom representations into an independently pretrained geometric latent diffusion model. Our adapters integrate pretrained all-atom representations through two mechanisms: enriching the diffusion model's molecular block features and incorporating pairwise structural information into the geometric diffusion model's denoiser. Theoretically, we prove that the augmented denoiser preserves geometry and pairwise attention fusion provides additional attention expressivity beyond feature-only fusion. Experimentally, across antibody, peptide, and small molecule binder-design benchmarks, PairFuse achieves stronger performance than the pretrained diffusion model and existing representation-fusion baselines.
The Smallest Grammar Problem (SGP) asks for a smallest straight-line grammar generating a given string, with size measured by the total right-hand-side length. For an input of length $n\ge2$ over at most $σ$ symbols, where $2\leσ\le n$, let $R=\log_σn$. We give a deterministic exact algorithm running in $n^{O(1)}2^{O(n\log\log n/(R\sqrt{\log n}))}$ time and polynomial space, with constants independent of $σ$. This gives time $2^{O(n\log\log n/(\log n)^{3/2})}$ for fixed alphabets and $2^{O(n\log\log n/\sqrt{\log n})}$ for arbitrary alphabets, up to polynomial factors. The algorithm combines pairwise disjoint representative occurrences of dictionary strings, a sparse interval certificate, and the Boolean tensor contraction theorem of OpenAI. It explicitly enforces dictionary membership, rule boundaries, and the grammar-size budget. Independently of that contraction theorem, disjoint representatives yield exact algorithms in $O^*(1.872258^n)$ time and $2^{O((n/R)\log(R+1))}$ time using $O(n)$ words of space. Anchor-based grammar projections establish binary NP-completeness and reductions with tunable alphabet size. We state their consequences as unconditional quantitative transfers from SGP algorithms to sparse SAT algorithms, and quantify the additional loss from sparsifying general formulas.
Recent work has explored optimizing LLM collaboration through Multi-Agent Reinforcement Learning (MARL), but many approaches still rely on predefined workflows, limiting efficiency and scalability. Decentralized LLM collaboration is challenging because agents execute independently with limited communication, share a joint objective, and must infer useful information from communication traces. In this paper, we propose BCMAAC, a Belief-Conditioned Multi-Agent Actor-Critic method that maintains beliefs over contextual information and uses them to facilitate training. Experiments on Minecraft game-playing, web searching and tool-using tasks show that BCMAAC improves task performance and collaborative behavior compared to naive actor-critic baselines. The improvements are particularly pronounced in multi-turn tasks with richer contextual information, where belief modeling provides a clear advantage.
Tianle Chen Xi'an Jiaotong University, Pengyu Cheng Xi'an Jiaotong University, Qiyuan Zhu, Jiacheng Wang Xi'an Jiaotong University, Bei Liu The Hong Kong University of Science and Technology, Hao Gu The Hong Kong University of Science and Technology, Ruijie Shen Toulouse School of Economics, Xiaofeng Hou Shanghai Jiaotong University 等 10 位作者
Large Language Models (LLMs) have achieved exceptional performance in complex reasoning via Chain-of-Thought (CoT), yet the associated computational costs remain prohibitive. CoT reasoning contains significant untapped efficiency potential across two dimensions: temporal redundancy, where reasoning steps may be unnecessary, and spatial redundancy, where computations can be performed at reduced precision. While current optimization techniques often necessitate resource-intensive fine-tuning or data curation, we introduce ASTRO (Adaptive Spatial and Temporal Redundancy Optimization), a training-free framework that simultaneously addresses both dimensions. ASTRO leverages Dewey’s reflective thinking model to segment reasoning phases, applying a progressive precision reduction strategy coupled with an entropy-based confidence mechanism for adaptive termination. Empirical results across diverse reasoning benchmarks demonstrate that ASTRO achieves up to an 11.3 \times efficiency gain without compromising accuracy, highlighting the advantages of holistic multi-dimensional redundancy management over isolated optimization methods.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Le Chen Argonne National Laboratory, Nuo Xu University of Minnesota - Twin Cities, Winson Chen University of Minnesota - Twin Cities, Bin Lei University of Minnesota - Twin Cities, Pei-Hung Lin Lawrence Livermore National Labs, Dunzhi Zhou University of Minnesota - Twin Cities and University of Florida, Rajeev Thakur Argonne National Laboratory, Caiwen Ding University of Minnesota - Twin Cities 等 10 位作者
Large language models (LLMs) have shown remarkable capabilities in code translation, yet their performance deteriorates in low-resource programming domains such as Fortran and emerging frameworks like CUDA, where high-quality parallel data are scarce. We present an automated dataset generation pipeline featuring a dual-LLM Questioner–Solver design that incorporates external knowledge from compilers and runtime feedback. Beyond traditional source–target code pair datasets, our approach additionally generates (1) verified translations with unit tests for assessing functional consistency, and (2) multi-turn dialogues that capture the reasoning process behind translation refinement. Applied to Fortran→C++ and C++→CUDA, the pipeline yields 3.64k and 3.93k dialogues, respectively. Fine-tuning on this data yields dramatic improvements in functional correctness, boosting unit test success rates by over 56% on the challenging C++-to-CUDA task. We show that the generated data enables a 7B open-weight model to significantly outperform larger proprietary systems on key metrics like compilation success.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Jiale Cheng, Yusen Liu, Xinyu Zhang, Yulin Fei The Hong Kong University of Science and Technology, Wenyi Hong Department of Computer Science and Technology, Tsinghua University, Ruiliang Lyu Zhipu AI, Weihan Wang, Zhe Su 等 14 位作者
Large language models (LLMs) conventionally represent text as sequences of discrete tokens, making long-context scaling largely a matter of processing more tokens more efficiently.We instead explore a complementary direction: increasing how much original context each token represents.To this end, we introduce Glyph, a framework that renders long texts into compact visual pages and processes them with a vision-language model (VLM), allowing a fixed context window to cover substantially more text.To make visual compression practical, Glyph combines continual pre-training on rendered long-text data, an LLM-driven genetic search to identify rendering configurations that balance compression and task performance, and post-training with supervised fine-tuning and reinforcement learning.Across multiple long-context benchmarks, Glyph achieves 3–4× token compression while maintaining performance comparable to strong text-only LLMs such as Qwen3-8B, with over 4× faster prefilling and decoding and 2× faster supervised fine-tuning.Under more aggressive compression, a VLM with a 128K context window can handle tasks that would otherwise require up to 1M input tokens.Our code and model are released at https://github.com/thu-coai/Glyph.
尚无 AI 解读，展示原文摘要
cs.CL · cs.CV
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Rishita Agarwal, Himanshu Singhal Indian Institute of Technology, Guwahati, Peter Baile Chen Massachusetts Institute of Technology, Manan Roy Choudhury Arizona State University, Dan Roth, Vivek Gupta Arizona State University
Answering natural language queries over relational data often requires retrieving and reasoning over multiple tables, yet most retrievers optimize only for query–table relevance and ignore table–table compatibility. We introduce REaR (Retrieve, Expand and Refine), a three-stage, LLM-free framework that separates semantic relevance from structural joinability for efficient, high-fidelity multi-table retrieval. REaR (i) retrieves query-aligned tables, (ii) expands these with structurally joinable tables via fast, precomputed column-embedding comparisons, and (iii) refines them by pruning noisy or weakly related candidates. Empirically, REaR is retriever-agnostic and consistently improves dense/ sparse retrievers on complex table QA datasets (BIRD, MMQA, and Spider) by improving both multi-table retrieval quality and downstream SQL execution. Despite being LLM-free, it delivers performance competitive with state-of-the-art LLM-augmented retrieval systems (e.g., ARM) while achieving much lower latency and cost. Ablations confirm complementary gains from expansion and refinement, underscoring REaR as a practical, scalable building block for table-based downstream tasks (e.g., Text-to-SQL).
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Large language models (LLMs) have demonstrated better safety performance in high-resource languages than in low-resource languages. We attribute this issue as a mismatch gap between language-agnostic semantic understanding ability and language dominant safety alignment biased toward high-resource languages. Based on above insights, we empirically identify the semantic bottleneck in LLMs: intermediate layers in which the geometry of model representations is governed primarily by shared semantic content rather than language identity. Then, we propose Language-Agnostic Semantic Alignment (LASA), which anchors safety alignment directly in semantic bottlenecks. Experiments show that LASA substantially improves safety across all languages: average attack success rate (ASR) drops from 24.7% to 2.8% on LLaMA-3.1-8B-Instruct and remains within 3–4% across Qwen2.5 and Qwen3 Instruct models (7B–32B). Besides, our analysis and method offer a representation-level perspective on LLM safety, suggesting that safety alignment requires anchoring safety understanding not in surface text, but in the model’s language-agnostic semantic space.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Data selection is a critical aspect of Reinforcement Learning with Verifiable Rewards (RLVR) for enhancing the reasoning capabilities of large language models (LLMs). Current data selection methods are largely heuristic-based, lacking theoretical guarantees and generalizability. This work proposes a theoretically-grounded approach using influence functions to estimate the contribution of each data point to the learning objective. To overcome the prohibitive computational cost of policy rollouts required for online influence estimation, we introduce an off-policy influence estimation method that efficiently approximates data influence using pre-collected offline trajectories. Furthermore, to manage the high-dimensional gradients of LLMs, we employ sparse random projection to reduce dimensionality and improve storage and computation efficiency. Leveraging these techniques, we develop Curriculum RL with Off-Policy Influence guidance (CROPI), a multi-stage RL framework that iteratively selects the most influential data for the current policy. Experiments on models up to 7B parameters demonstrate that CROPI significantly accelerates training. On a 1.5B model, it achieves a 2.66x step-level acceleration while using only 10% of the data per stage compared to full-dataset training. Our results highlight the substantial potential of influence-based data selection for efficient RLVR.
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表
Erel Kaplan, Tomer Bitan, Lian Ghrayeb, Le Chen Argonne National Laboratory, Tom Yotam Code Metal, Niranjan Hasabnis Codemetal AI, Gal Oren Stanford University and Computer Science Department, Technion - Israel Institute of Technology
Parallel programming is central to HPC and AI, but producing code that is correct and fast remains challenging, especially for OpenMP GPU offload, where data movement and tuning dominate. Autonomous coding agents can compile, test, and profile on target hardware, but outputs are brittle without domain scaffolding.We present ParaCodex, an HPC-engineer workflow that turns a Codex-based agent into an autonomous OpenMP GPU offload system using staged hotspot analysis, explicit data planning, correctness gating, and profiling-guided refinement. We evaluate translation from serial CPU kernels to OpenMP GPU offload kernels on HeCBench, Rodinia, and NAS. After excluding five kernels, ParaCodex succeeded on all 31 valid kernels. In 27/31 (87%) of these valid cases, the generated kernels improved GPU time over reference implementations, a result that holds independently on both the A100 and RTX 4060. The resulting OpenMP kernels achieve geometric-mean speedups of 3.1 (A100) and 3.6 (RTX 4060) on HeCBench and 1.5 and 1.1 on Rodinia, and outperform a zero-shot Codex baseline on all suites. We also evaluate CUDA -> OpenMP offload translation on ParEval, where ParaCodex maintains high compilation and validation rates in code-only and end-to-end settings.ParaCodex is available at https://github.com/Scientific-Computing-Lab/ParaCodex
尚无 AI 解读，展示原文摘要
cs.CL
Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) · 2026 · 正式发表