The Mirage of Optimizing Training Policies: Monotonic Inference Policies as the Real Objective for LLM Reinforcement Learning
Paper • 2606.29526 • Published • 161
Mason Baker
masonbaker5
·
AI & ML interests
None yet
Recent Activity
liked a model about 8 hours ago
deepseek-ai/DeepSeek-OCR liked a Space 3 days ago
AimeeBingmouQu/ProtectBirds liked a model 4 days ago
mradermacher/Peaceful-Days-12B-i1-GGUFOrganizations
None yet
upvoted 2
papers 5 days ago
upvoted 4
papers about 2 months ago
Full Attention Strikes Back: Transferring Full Attention into Sparse within Hundred Training Steps
Paper • 2605.16928 • Published • 98
SafeDiffusion-R1: Online Reward Steering for Safe Diffusion Post-Training
Paper • 2605.18719 • Published • 7
CiteVQA: Benchmarking Evidence Attribution for Trustworthy Document Intelligence
Paper • 2605.12882 • Published • 274
Mean Mode Screaming: Mean--Variance Split Residuals for 1000-Layer Diffusion Transformers
Paper • 2605.06169 • Published • 238
upvoted 5
papers 3 months ago
Rethinking Generalization in Reasoning SFT: A Conditional Analysis on Optimization, Data, and Model Capability
Paper • 2604.06628 • Published • 329
Adam's Law: Textual Frequency Law on Large Language Models
Paper • 2604.02176 • Published • 509
ClawKeeper: Comprehensive Safety Protection for OpenClaw Agents Through Skills, Plugins, and Watchers
Paper • 2603.24414 • Published • 183
FIPO: Eliciting Deep Reasoning with Future-KL Influenced Policy Optimization
Paper • 2603.19835 • Published • 353
6Bit-Diffusion: Inference-Time Mixed-Precision Quantization for Video Diffusion Models
Paper • 2603.18742 • Published • 11