Researchers say Anthropic confirmed its models were distilled the way their paper described
HildeKuehne · x · 2026-09-11
Researcher @kotekjediml says they finally got confirmation from Anthropic that its models were indeed distilled in the way described in their paper. An Anthropic researcher replied: "we've seen so much traffic that uses attacks you described in that paper" — indirectly corroborating the distillation claim.
More from Models
- V4.1 model reportedly has a flaky January 2026 knowledge cutoff — teortaxesTex · 2026-09-12
- Dev Claims DeepSeek v4.1 Flash "Way Better" Than Gemini Flash Models — gaganghotra_ · 2026-09-12
- Altman Talks GPT7 in New Video Clip — Popular_Slip_5311 · 2026-09-12
- LinkedIn user claims GPT-6 built a pixel-perfect Figma design system in 3 hours — AIandDesign · 2026-09-12
- Report: Google Achieved Recursive Self-Improvement in Next-Gen AI Model, Launching Oct 5 — Dr_Singularity · 2026-09-12
- Leaked Codenames Suggest OpenAI Model Tiers Map to Claude's Opus/Sonnet/Haiku — haider1 · 2026-09-12