John Schulman releases Inkling 2 Small, detailing alignment surprises in training
corbtt · x · 2026-11-06
John Schulman announces the release of Inkling 2 Small, accompanied by a blog post with extensive detail on the training process.
He says the most interesting part of the release is what the team learned about alignment (see the "Safety and alignment" and "Training Inkling 2 Small" sections): early checkpoints exhibited some unpleasant, unexpected behaviors that informed their approach.
More from Models
- Post-trained 35B Qwen learns weight-space programming: Freeway score jumps 21.0→47.1 without training on it — CatAstro_Piyush · 2026-11-07
- Claude now casually generates design concept visuals while working, user delighted — RileyRalmuto · 2026-11-07
- New AI lab reportedly squeezes more from 12T tokens than DeepSeek gets from 45T — teortaxesTex · 2026-11-07
- User shares approval email for Anthropic's Cyber Verification Program Defense Access — AIFlow_ML · 2026-11-07
- As ChatGPT ends unlimited chat, Plus subscribers weigh canceling and moving to Claude — ShinyPlasm · 2026-11-07
- Qwen 4 Preview Appears on Arena, Qwen 4 Max Rumored With 1M-Token Context — WorldofAI · 2026-11-07