paperHugging Face Daily Papers·21 Oct 20263.3
TraceForge distills with the student own reasoning traces.
paperarXiv (systems, data, ML)·23 Oct 20263.3
Traces membership imprints in on-policy self-distillation.
paperarXiv (systems, data, ML)·23 Oct 20263.3
Advantage-based rule for when on-policy distillation should follow the teacher.
paperHugging Face Daily Papers·19 Oct 20263.3
Bridged self-distillation for on-policy learning.
paperIEEE Xplore·1 Nov 20263.1
Distills reliable recommendation rationales from LLMs for conversational recommenders.