Researchers at ACL 2026 introduced a method to reuse LLM hidden states for classification tasks, training lightweight probes on token- and layer-selective representations to eliminate separate guard m…
Researchers from multiple institutions published a comprehensive survey on Process Reward Models (PRMs) at ACL 2026, covering data generation, model construction, and usage for step-level reasoning ev…
Researchers Bhavik Chandna and Procheta Sen introduced a counterfactual explanation framework for retrieval models that identifies which terms need to be added to a document to improve its ranking, ad…
Researchers at the 64th Annual Meeting of the Association for Computational Linguistics introduced a dual-task paradigm combining arithmetic and sentence comprehension to study how language models all…
Researchers at ACL 2026 found that entropy collapse in LLM reasoning, which undermines test-time scaling, is driven by premature overconfidence at a small set of critical tokens. They proposed SCOPE, …
Researchers at French institutions found that diversity-driven sampling can reduce pre-training dataset size by up to 94% and training time by 73% while maintaining performance in ModernBERT models. I…
Researchers at ACL 2026 revealed PriceBlind, a visual adversarial attack that exploits a vulnerability called Visual Dominance Hallucination in multimodal large language models used as financial agent…
Researchers at the Association for Computational Linguistics found that large language models' numerical accuracy drops significantly when numbers are presented in underrepresented scripts or formats,…
Researchers introduced CoLabScience, a proactive AI assistant that uses a reinforcement learning framework called PULI to autonomously intervene in biomedical research discussions. The system, trained…
Researchers at Northwestern University and the University of Chicago developed a computational method to measure open codes in qualitative analysis, using LLM-enriched hierarchical clustering and four…
Researchers introduced PARAMΔ, a method that upcycles a dense LLM into a Mixture-of-Experts (MoE) architecture to expand language capabilities without costly alignment. By grafting a MoE-expanded para…
Researchers at the Association for Computational Linguistics introduced a Dual-Phase Self-Evolution (DPSE) framework for large language models that jointly optimizes user preference adaptation and dom…
Researchers introduced DySECT, a dynamic self-evolving extraction system that uses a closed-loop cycle where an LLM extracts triples to populate a knowledge base, which then improves the LLM via promp…