LLM-guided refinement of Python policies reaches competitive Atari scores with fewer environment interactions than standard deep RL, on Pong, Breakout, and Space Invaders.
Playing nethack with llms: Potential & limitations as zero-shot agents
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Learning Game-Playing Agents with Generative Code Optimization
LLM-guided refinement of Python policies reaches competitive Atari scores with fewer environment interactions than standard deep RL, on Pong, Breakout, and Space Invaders.