reasoning-degeneration-dev/

gepa-abstract-reasoning-vanilla

GEPA prompt optimization experiment on AIME math problems. Task LM: openai/gpt-4.1-mini | Reflection LM: openai/gpt-5 | Reflection Mode: abstract_reasoning | Last updated: 2026-02-27 01:21 UTC

	Results




	

Run Method k Mode Val Score Test Acc Tokens Cost Time

fixed_vanilla_k3 vanilla 3 abstract_reasoning 31.11% N/A 55,714 $0.1350 2185s

	Learning Curves





	
	Experiment Config

{ "script_name": "run_experiment.py"… See the full description on the dataset page: https://huggingface.co/datasets/reasoning-degeneration-dev/gepa-abstract-reasoning-vanilla.