WTF GENIUS PAPERS Collection Papers that made me appreciate my major and my life a little more. obs=Observation, innov=Innovation. Most papers are abt improving tiny models. • 298 items • Updated 1 day ago • 69
FlashPrefill V2: Block-Sparse Prefill Attention for Long-Context LLM Serving Paper • 2608.19758 • Published 2 days ago • 14
FACET: Preserving Source Intent and Executable State in Terminal Task Synthesis Paper • 2608.18580 • Published 4 days ago • 111
Inject, Align, Recover: Staged Post-Training for Retrieval-Free Document Knowledge Internalization Paper • 2608.20281 • Published 3 days ago • 10
WTF GENIUS PAPERS Collection Papers that made me appreciate my major and my life a little more. obs=Observation, innov=Innovation. Most papers are abt improving tiny models. • 298 items • Updated 1 day ago • 69
LLMs Get Smarter from Targeted Synthetic Multilingual Data Paper • 2608.15964 • Published 7 days ago • 4
EnvHarness: Awakening Static Worlds for Agent Learning Paper • 2608.19880 • Published 3 days ago • 245
Thinking in a Low-Resource Language: What SFT Builds, What RL Fixes, What Accuracy Cannot See Paper • 2608.17744 • Published 4 days ago • 13 • 3
WTF GENIUS PAPERS Collection Papers that made me appreciate my major and my life a little more. obs=Observation, innov=Innovation. Most papers are abt improving tiny models. • 298 items • Updated 1 day ago • 69
Thinking in a Low-Resource Language: What SFT Builds, What RL Fixes, What Accuracy Cannot See Paper • 2608.17744 • Published 4 days ago • 13
FreeToken: Efficient Edge-Native MoE Serving with Bandwidth-Adaptive Execution Paper • 2608.16157 • Published 6 days ago • 76
WTF GENIUS PAPERS Collection Papers that made me appreciate my major and my life a little more. obs=Observation, innov=Innovation. Most papers are abt improving tiny models. • 298 items • Updated 1 day ago • 69
MathForm: Scaling Mathematical Autoformalization with Knowledge Retrieval and Verification-Guided Refinement Paper • 2608.14221 • Published 8 days ago • 8
SPADE: Self-Play in Adaptive Synthetic Executable Environments Paper • 2608.19197 • Published 4 days ago • 47
WTF GENIUS PAPERS Collection Papers that made me appreciate my major and my life a little more. obs=Observation, innov=Innovation. Most papers are abt improving tiny models. • 298 items • Updated 1 day ago • 69
Co-RL: Unsupervised Reasoning Emerges from Diverse Cohort in Multi-agent RL Paper • 2608.17253 • Published 4 days ago • 91
WTF GENIUS PAPERS Collection Papers that made me appreciate my major and my life a little more. obs=Observation, innov=Innovation. Most papers are abt improving tiny models. • 298 items • Updated 1 day ago • 69
Looped Language Models Improve Compositional Tool Calling Paper • 2608.18171 • Published 6 days ago • 16
StateM: Reaching 95.3% Raw Accuracy, or a \$15 Frontier Run, on Terminal-Bench 2.1 via Harness Scaling Paper • 2608.15089 • Published 8 days ago • 435
Demystifying Agent Skills: Why They Work-Until They Don't Paper • 2608.14036 • Published 9 days ago • 162