Steered-teacher self-distillation cuts reference hallucinations in thinking models
Activation steering is applied only to the teacher; the student is not steered, and AIME and HMMT accuracy rises over vanilla OPSD.
Stories per day · 30 days
2 stories, newest first · one card per story
Stories tagged #hmmt
From Pareto to Preference: Personalized Test-Time Scaling via Amortized Agentic Policy Discovery
PersonTTS reuses prior search to find test-time controllers matching each user's accuracy, latency, and cost goals.