Benchmarks and evaluation

Leaderboards, eval harnesses and ablations — the contested business of deciding which model is actually better.

23,273 stories · RSS feed

arXiv AI

Jun 9

Beyond Pass Rate: A Multilingual, Execution-Grounded Evaluation of Open Code LLMs

arXiv:2606. 08840v1 Announce Type: new Abstract: Code generation models are typically compared using compact execution benchmarks and aggregate pass rates, but such summaries obscure how performance varies across programming languages, problem families, and failure modes.

By Sayed Erfan Arefin

llms benchmarks

arXiv AI

Jun 9

ConflictRAG: Detecting and Resolving Knowledge Conflicts in Retrieval Augmented Generation

arXiv:2605. 17301v2 Announce Type: replace-cross Abstract: Retrieval-Augmented Generation (RAG) systems implicitly assume mutual consistency among retrieved documents -- an assumption that frequently fails in practice.

By Chenyu Wang, Yueyuan Li, Yingmin Liu, Yang Shu

llms rag benchmarks

arXiv AI

Jun 9

InA-Probe: Instruction-Aware Active Probing for Time Series Forecasting with LLMs

arXiv:2606. 08601v1 Announce Type: new Abstract: Large Language Models (LLMs) have recently demonstrated impressive potential for time series forecasting.

By Peiliang Gong, Emadeldeen Eldele, Chenyu Liu, Ziyu Jia, Yi Ding, Xinliang Zhou, Lianchao Gu, Qi Zhu, Yang Liu, Daoqiang Zhang, Xiaoli Li

llms benchmarks safety

arXiv AI

Jun 9

Riemannian-Manifold Steering: Geometry-Aware Generative Autoencoders for Label-Free Steering

arXiv:2605. 24942v2 Announce Type: replace-cross Abstract: Steering a language model - intervening on its internal activations to change downstream behaviour - has recently expanded beyond linear interpolation to nonlinear methods such as angular and kernelized steering, which define intervention transformations without learning an explicit geometry over paths in activation space.

By Narmeen Oozeer, Shivam Raval, Philip Quirke, Manikandan Ravikiran, Jeff Phillips, Shriyash Upadhyay, Amirali Abdullah

llms benchmarks

arXiv AI

Jun 9

SecureVibeBench: Benchmarking Secure Vibe Coding of AI Agents via Reconstructing Vulnerability-Introducing Scenarios

arXiv:2509. 22097v5 Announce Type: replace-cross Abstract: Large language model-powered code agents are rapidly transforming software engineering, yet the security risks of their generated code have become a critical concern.

By Junkai Chen, Huihui Huang, Yunbo Lyu, Junwen An, Jieke Shi, Chengran Yang, Ting Zhang, Haoye Tian, Yikun Li, Zhenhao Li, Xin Zhou, Xing Hu, David Lo

llms agents benchmarks

arXiv AI

Jun 9

Executable World Models for ARC-AGI-3 in the Era of Coding Agents

arXiv:2605. 05138v2 Announce Type: replace Abstract: We evaluate an initial coding-agent system for ARC-AGI-3 in which the agent maintains an executable Python world model, verifies it against previous observations, refactors it toward simpler abstractions as a practical proxy for an MDL-like simplicity bias, and plans through the model before acting.

By Sergey Rodionov

llms agents benchmarks safety

arXiv AI

Jun 9

Subtitle-Aligned Fine-Tuning of Whisper for Swiss German ASR: Benchmark Contamination, Convention Mismatch, and an Honest Baseline at 25.6% WER (13.8% cWER)

arXiv:2606. 07608v1 Announce Type: cross Abstract: We present a systematic study of fine-tuning OpenAI's Whisper large-v3 for Swiss German ASR, using 1,367 hours of broadcast speech paired with Standard German subtitles as weak supervision.

By Felix Akeret

fine-tuning multimodal benchmarks safety data-engineering

arXiv Machine Learning

Jun 9

Helpful to a Fault: Measuring Illicit Assistance in Multi-Turn, Multilingual LLM Agents

arXiv:2602. 16346v4 Announce Type: replace-cross Abstract: LLM-based agents execute real-world workflows via tools and memory.

By Nivya Talokar, Ayush K Tarun, Murari Mandal, Maksym Andriushchenko, Antoine Bosselut

llms agents benchmarks safety

arXiv Machine Learning

Jun 9

When Do Local Score Models Extrapolate Across Size? A Diagnostic Theory and Benchmark

arXiv:2606. 09705v1 Announce Type: new Abstract: Scientific generative modeling often requires size transfer, where models trained on small systems are evaluated on larger ones.

By Wenjie Xi

diffusion benchmarks

arXiv AI

Jun 9

RiskNet: A large-scale dataset of AI risk incidents from news with alignment and multi-dimensional annotations

arXiv:2606. 08376v1 Announce Type: cross Abstract: As artificial intelligence (AI) systems are increasingly deployed across socially consequential domains, reports of AI-related harms and failures have grown in frequency and diversity.

By Leihan Zhang, Wecheng Ye, Xianlong Ma, Haochuan Liu, Yang Li, Qianyu Zhang, Jinliang Chen, Qiang Yan

benchmarks safety

arXiv AI

Jun 9

IMUG-Bench: Benchmarking Unified Multimodal Models on Interleaved Understanding and Generation

arXiv:2606. 09169v1 Announce Type: new Abstract: In recent years, unified multimodal models (UMMs) have emerged to support both understanding and generation within a single framework.

By Lingyi Meng, Zecong Tang, Haoran Li, Tengju Ru, Zhejun Cui, Weitong Lian, Qi Kang, Hangshuo Cao, Yichen Zhu, Yechi Liu, Kaixuan Wang, Yu-Jie Yuan, Chunwei Wang, Yu Zhang, Bo Dai

multimodal benchmarks safety

arXiv AI

Jun 9

TheoremBench: Evaluating LLMs on Theorem Proving in Formal Mathematics

arXiv:2606. 09450v1 Announce Type: new Abstract: LLMs have recently achieved strong results on formal proving benchmarks.

By QuocViet Pham, Elvir Karimov, Andrey Galichin, Ivan Oseledets

llms benchmarks

arXiv AI

Jun 9

Structure-Conditioned Actor-Critic Branches for Quality-Diversity Reinforcement Learning

arXiv:2606. 08735v1 Announce Type: new Abstract: Quality-diversity reinforcement learning (QD-RL) aims to construct policy repertoires that contain both high-performing and behaviorally diverse policies.

By Lianrong Zuo, Peilan Xu, Yong Liu, Wenjian Luo

reinforcement-learning efficiency benchmarks

arXiv Machine Learning

Jun 9

UNIQ: Conformal Calibration for Adaptive Conservatism in Offline Reinforcement Learning

arXiv:2606. 07592v1 Announce Type: new Abstract: Offline reinforcement learning requires careful conservatism to mitigate distribution shift, yet most existing methods apply a fixed penalty uniformly across all states regardless of local data coverage.

By Aditya Upadhyay

reinforcement-learning benchmarks

arXiv Machine Learning

Jun 9

Optimizing Few-Step Generation with Adaptive Matching Distillation

arXiv:2602. 07345v2 Announce Type: replace-cross Abstract: Distribution Matching Distillation (DMD) is a powerful acceleration paradigm, yet its stability is often compromised in Forbidden Zone, regions where the real teacher provides unreliable guidance while the fake teacher exerts insufficient repulsive force.

By Lichen Bai, Zikai Zhou, Shitong Shao, Wenliang Zhong, Shuo Yang, Shuo Chen, Bojun Chen, Zeke Xie

diffusion efficiency benchmarks

arXiv AI

Jun 9

Post-training is (Massive) Supervised Learning

arXiv:2606. 07527v1 Announce Type: cross Abstract: The prevailing paradigm for training LLMs has evolved to rely on a massive post-training phase consisting of SFT and RL.

By Michael Hassid, Yossi Adi, Roy Schwartz

llms reinforcement-learning fine-tuning benchmarks

arXiv Machine Learning

Jun 9

Escaping the KL Agreement Trap in On-Policy Distillation

arXiv:2606. 09471v1 Announce Type: new Abstract: On-policy distillation (OPD) provides dense token-level supervision by asking a teacher to score student-generated rollouts.

By Haoran Xin, Anhao Zhao, Ying Sun, Jin Li, Xiaoyu Shen, Hui Xiong

efficiency benchmarks

arXiv AI

Jun 9

A Multi-modal Agentic Co-pilot for Evidence Grounded Computational Pathology

arXiv:2606. 08093v1 Announce Type: new Abstract: Pathology is the cornerstone of modern medicine, where accurate decision-making relies heavily on evidence-based practices.

By Zhe Xu, Zhengyu Zhang, Zhiyuan Cai, Jiahao Xu, Yijie Lin, Ziyi Liu, Junlin Hou, Hongyi Wang, Yuxiang Nie, Ling Liang, Yihui Wang, Yingxue Xu, Ronald Cheong Kin Chan, Li Liang, Hao Chen

agents multimodal benchmarks

arXiv Machine Learning

Jun 9

Few-step Cofolding with All-Atom Flow Maps

arXiv:2606. 08375v1 Announce Type: new Abstract: All-atom generative modeling of 3D biomolecular complexes has emerged as the dominant paradigm for predicting the structure of proteins and protein-ligand systems.

By Gianluca Scarpellini, Ron Shprints, Peter Holderrieth, Juno Nam, Pranav Murugan, Rafael G\'omez-Bombarelli, Tommi Jaakola, Maruan Al-Shedivat, Nicholas Matthew Boffi, Avishek Joey Bose

diffusion efficiency benchmarks safety

arXiv AI

Jun 9

Skill Retrieval Augmentation for Agentic AI

arXiv:2604. 24594v3 Announce Type: replace-cross Abstract: As large language models (LLMs) evolve into agentic problem solvers, they increasingly rely on external, reusable skills to handle tasks beyond their native parametric capabilities.

By Weihang Su, Jianming Long, Qingyao Ai, Qiaozhi He, Yichen Tang, Changyue Wang, Yiteng Tu, Yingbo Wang, Yiqun Liu

llms agents benchmarks

Get the daily briefing