Why LLMs Struggle with Abstract Meaning Comprehension
Explore why large language models struggle with abstract meanings and how fine-tuning and new methods improve their comprehension abilities.
BayMOTH: Efficient Bayesian Optimization with Meta-Lookahead
Discover BayMOTH, a novel Bayesian optimization method that boosts sample efficiency using meta-lookahead and adaptive strategies for AI tasks.
NTIRE 2026 Cross-Domain Few-Shot Object Detection Results
Explore top methods and results from the NTIRE 2026 challenge on cross-domain few-shot object detection with limited data and domain shifts.
Filtered Reasoning Score: Assessing LLM Reasoning Quality
Discover the Filtered Reasoning Score (FRS) to evaluate large language models' reasoning quality beyond accuracy using top-confident traces.
INDOTABVQA: Cross-Lingual Table VQA Benchmark for Bahasa
Explore INDOTABVQA, a benchmark for cross-lingual table understanding in Bahasa Indonesia documents, enhancing Vision-Language Models' accuracy.