SWE-RL: Advancing LLM Reasoning via Reinforcement Learning on Open Software Evolution Paper β’ 2502.18449 β’ Published 1 day ago β’ 34
The Relationship Between Reasoning and Performance in Large Language Models -- o3 (mini) Thinks Harder, Not Longer Paper β’ 2502.15631 β’ Published 5 days ago β’ 6
The Ultimate Collection of Code Classifiers Collection π₯ 15 classifiers, 124M parameters, one per programming languageβ for assessing the educational value of GitHub code β’ 15 items β’ Updated 6 days ago β’ 10
Native Sparse Attention: Hardware-Aligned and Natively Trainable Sparse Attention Paper β’ 2502.11089 β’ Published 11 days ago β’ 134
SWE-Lancer: Can Frontier LLMs Earn $1 Million from Real-World Freelance Software Engineering? Paper β’ 2502.12115 β’ Published 9 days ago β’ 41
view reply We have also had good success applying tts more broadly at a diverse set of tasks in optillm - https://github.com/codelion/optillm