Skill Reward Hacking
Fresh
Reward Hacking Sprint v3: Enhanced environment with true metrics tracking, harder proxy traps, and hacking detection. 12 proxy rewards (4 traps) + ...
- Type
- RL Env
- Tags
- Reward HackingText GenerationSkill CreationSmall ModelsHacking DetectionHidden RewardDashboardCurriculumSimple ModeScientific LeversEarly WarningHack TaxonomyTransfer MatrixCorrectnessProxy Vs TrueCross EvalOnset Prediction
- Runtime
multi-turn- License
- mit
- Size
- v10.3.1
- Published
- May 2026
- Updated
- Sep 2026
Cite
Notes
Only stored in your browser.
Public scores on this env
44 vf-eval reports across 4 models
1MiMo-V2.5-ProXiaomi9.482Llama 3.2 Instruct 1BMeta Platforms8.963GLM 5.1Zai5.244GPT-5 MiniOpenAI4.95
Open the scoring view →