1/ 🎮 Can training on games improve LLM reasoning?
Introducing GlyphBench: 360+ game tasks for language-model RL.
We use it as a benchmark, a research tool, and a post-training suite.
🧵⬇️
astra playing NetHack:
- score: 48,978
- BALROG progress: 49.4%
- max dungeon depth: 17
- max experience level: 14
important note: this is a cherry-picked episode (best of 10), and Astra has been able to modify its own game-playing harness
no privileged information or unfair
You should try Silico. Not only for mech interp, but for AI Research in general.
Goodfire have put a lot work into building a great interface for long-horizon agent-driven research.
Silico, the platform for ambitious AI research, is publicly available today.
AI is advancing fast. The tools to understand it need to advance even faster. Silico lets you interpret and train your models at frontier scale.
Learn more + get access 🧵
Training a model to generate RL tasks not too hard, not too easy costs many solver runs per task.
PROPEL predicts difficulty via a probe on its activations instead, amortizing cost and speeding up generator optimization.
New open-ended RL research from @vmax + @GoodfireAI.
