Researcher proposes calibrating CoT disclosure: distillation efficiency tracks disclosed reasoning
eliebakouch · x · 2026-10-18
Elie Bakouch hypothesizes that distillation effectiveness (information gained per sample) correlates predictably with how much reasoning is disclosed to users—you could train models at different disclosure levels, measure samples needed, and calibrate what to show. Open questions: whether minimal reasoning info suffices, and whether fixing this at the model level creates a deeper tradeoff between CoT controllability and distillation resistance.
Related event: Stolen Thoughts paper sparks debate over distillation and CoT theft(9 posts)→
More from Models
- LLMs used to be wrong, now they're bullshit, says Konrad Kording — KordingLab · 2026-10-19
- Pi harness nearly matches DeepSeek's own at 42.2 vs 42.5 on Bug Hunt Bench, at half the cost — PawelHuryn · 2026-10-19
- Krea2 uncensoring tools vanish: repo down, dev banned, refusal-reduction LoRAs removed — softlarch · 2026-10-19
- 'I miss 5.5': users say new models are better but lost their personality — Dapper_Ambition5963 · 2026-10-19
- Anthropic says dots will keep working in the EU and launch there later — Dimillian · 2026-10-18
- Gemini Fails to Identify the Prime Minister in Latest Capability Slip — LemonTeaCake007 · 2026-10-18