📰 AI News — Latest AI Updates

Page 755 News

LongSumEval: Reshaping Long-Document Summarization Evaluation with QA Feedback

A latest arXiv paper introduces LongSumEval, a framework that unifies summarization evaluation and g...

research 24 2026-04-29

New Research Reveals Hidden Mental Health Discrimination in LLM Reasoning

A new study analyzing the intermediate reasoning steps of large language models reveals hidden biase...

research 24 2026-04-29

Dual-Track CoT: Enabling Efficient Reasoning for Small Language Models

A latest arXiv paper proposes the Dual-Track CoT method, which uses a budget-aware stepwise guidance...

research 21 2026-04-29

Why Does Reinforcement Learning Generalize? Feature-Level Mechanistic Study Reveals Secrets of LLM Post-Training

A latest arXiv paper analyzes feature-level mechanisms to reveal why reinforcement learning post-tra...

research 24 2026-04-29

"Roundtrip Verification" Makes LLM Autoformalization More Faithful and Reliable

A new study proposes a label-free "roundtrip verification and repair" method that effectively detect...

research 20 2026-04-29

Survey Paper Analysis: A Panoramic View of LLM-Driven Conversational User Simulation Research

A latest arXiv survey systematically reviews research progress in LLM-based conversational user simu...

research 22 2026-04-29

LiteLLM Hit by Critical SQL Injection Vulnerability, Exploited in the Wild Within 36 Hours

BerriAI's open-source project LiteLLM has been found to contain a critical SQL injection vulnerabili...

industry 22 2026-04-29

BenchGuard: Using AI to Audit AI Benchmarks

A research team introduces the BenchGuard framework, the first to leverage frontier large language m...

research 21 2026-04-29

GAIA-v2-LILT: A Multilingual AI Agent Benchmark That Goes Beyond Translation

Researchers introduce GAIA-v2-LILT, a refined pipeline combining functional alignment and cultural a...

research 18 2026-04-29

ADE Adaptive Dictionary Embeddings: Breaking Through the Word Representation Bottleneck in Large Language Models

Researchers propose ADE (Adaptive Dictionary Embeddings), the first method to successfully extend mu...

research 18 2026-04-29

Independent Component-Based Brain Encoding Model Breaks Through fMRI Research Bottleneck

Researchers propose an independent component (IC)-based brain encoding framework that effectively se...

research 21 2026-04-29

When AI Meets Classic Film Aesthetics: Starting From a Coat

Starting from the iconic coat in the classic film Withnail & I, the community debates AI's role in r...

opinion 20 2026-04-29