🏢 Renmin University of China

OlymMATH: A new Olympiad-level math benchmark rigorously tests LLMs’ reasoning, revealing limitations and paving the way for advancements.

ETVA evaluates text-to-video alignment via fine-grained question generation and answering.

MathFusion: Instruction Fusion enhances LLM’s math problem-solving!

PLM retrievers overrate low-perplexity docs, causing source bias. This paper reveals the causal effect & offers a fix!

SEAP: Unlock LLM brainpower w/ training-free sparse expert activation pruning! Boost efficiency, keep accuracy. Optimize LLMs now!

eMIGM: A unified, efficient masked image generation model achieving state-of-the-art performance with fewer resources.

R1-Searcher: RL enhances LLMs by incentivizing autonomous search, outperforming RAG methods, even GPT-4o-mini!

This paper explores and improves R1-like reasoning models through RL and tool manipulation, achieving significant accuracy gains.

SURVEYX automates academic survey generation, enhancing content and citation quality.

YuLan-Mini: An open, data-efficient 2.42B parameter LLM achieving top-tier performance with innovative training techniques.

RetroLLM unifies retrieval & generation in LLMs, boosting accuracy and cutting costs.

HtmlRAG boosts RAG system accuracy by using HTML, not plain text, to model retrieved knowledge, improving knowledge representation and mitigating LLM hallucination.