Rethinking Token Reweighting for SFT: Suppress, Reverse, and Extrapolate Learned Features
Paper • 2609.33463 • Published • 11
Ctrl+Alt+Del
mikectrlaltdel
·
AI & ML interests
None yet
Recent Activity
upvoted a paper about 18 hours ago
Rethinking Token Reweighting for SFT: Suppress, Reverse, and Extrapolate Learned Features upvoted a paper about 18 hours ago
Spatial Memory Intelligence: Endowing World Models with Understanding-Driven Long-Term Memory upvoted a paper about 18 hours ago
RealCompanion: Benchmarking Human Understanding from Reasoning over Longitudinal Real-World ConversationsOrganizations
None yet
upvoted 3
papers about 18 hours ago
liked a
model 2 days ago
upvoted 3
papers 2 days ago
Prefill-Free Cross-Family KV Cache Transfer for Heterogeneous Multi-Agent LLMs
Paper • 2609.32259 • Published • 86
Predictive Credit: Measuring What Scientific Explanations Add to Experimental Forecasts
Paper • 2610.00314 • Published • 100
ActiveSaddler: Automated Curriculum Learning for Agent Harness Optimization
Paper • 2610.00906 • Published • 74