List your company

Logical Reasoning on GeoShape BBH

90AccuracyMemAPO
20.3238.4156.574.59Mar 23, 2026

Evaluation results

MethodLinks
MemAPOBackbone=GPT-4o-mini2026.0390
MemAPOBackbone=Qwen3-8B2026.0377
TextGradBackbone=GPT-4o-mini2026.0376
ProTeGiBackbone=GPT-4o-mini2026.0371
CoTBackbone=Qwen3-8B2026.0364
Step-BackBackbone=GPT-4o-mini2026.0361
Step-BackBackbone=Qwen3-8B2026.0358
ProTeGiBackbone=Qwen3-8B2026.0356
SPOBackbone=GPT-4o-mini2026.0354
RaRBackbone=Qwen3-8B2026.0353
CoTBackbone=GPT-4o-mini2026.0347
RaRBackbone=GPT-4o-mini2026.0343
PromptBreederBackbone=GPT-4o-mini2026.0342
TextGradBackbone=Qwen3-8B2026.0339
OPROBackbone=Qwen3-8B2026.0336
IOBackbone=GPT-4o-mini2026.0335
OPROBackbone=GPT-4o-mini2026.0335
SPOBackbone=Qwen3-8B2026.0328
IOBackbone=Qwen3-8B2026.0323
PromptBreederBackbone=Qwen3-8B2026.0323