policy
PolicyCliff
SafeAGI-01 · PyTorch
or hover any field below to flag it
Overview
Name
PolicyCliff
Author
SafeAGI-01
Framework
PyTorch
License
unknown
Skill type
navigation
Evidence level
untested
Task description
The Policy Cliff: A Theoretical Analysis of Reward-Policy Maps in Large Language Models. A rigorous mathematical framework analyzing the stability of the reward–policy mapping in RL-trained LLMs.
Spaces
Action space
other · 0-dim · 0Hz
Observation space
- type: other
Links
HuggingFace repo
null
Paper (arXiv)
null
Compatible robots
3+17 mentioned but not in catalog yetCompatible environments
0No environments list PolicyCliff yet.
Datasets that reference this policy
0No datasets reference PolicyCliff yet.