Astra's perception got much better: mostly scale plus pretraining encoder wins
willdepue · x · 2026-09-07
In the ongoing discussion on Astra's 3D modeling leap, willdepue notes the model's perception improved dramatically — "probably mostly just scale, but I'm sure some pretraining arch/encoder win." Astra "can see really well and this surely helps." Combined with his earlier take, he attributes the overall jump largely to better-targeted RL environments.
Related event: Debate Over Astra's 3D Leap: No-CoT Ability or Scale?(3 posts)→
More from Models
- User catches model failing to parse 400-paper file, citing stale news sources instead — jungofthewon · 2026-10-27
- Odd eval result: Gemini 4.1 Flash gets more accurate when asked to control its CoT — dfrsrchtwts · 2026-10-27
- Poster surprised that v4.1 flash gets more accurate when asked to control its CoT — dfrsrchtwts · 2026-10-27
- Dev claims Anthropic is baking 'sabotage' into model behavior for 'wrong' engineering — cephaloform · 2026-10-27
- Gemini fails at listing unread emails, tells user to search Gmail manually — gethackteam · 2026-10-27
- Meta's Muse Spark Open goes live on Applied Compute for async RL post-training — ypatil125 · 2026-10-27