List your company

Text-based Game Playing on Jericho (Test-Time Learning)

95Detective ScoreEvoTest
18.0438.025877.98Oct 15, 2025

Evaluation results

MethodLinks
EvoTestBackbone=anthropic/claude-4-sonnet-20250522 (C)2025.1095801638353450
EvoTestBackbone=google/gemini-2.5-flash (G)2025.1094771435323147
EvoPromptBackbone=anthropic/claude-4-sonnet-20250522 (C)2025.1067501232262936
PromptbreederBackbone=anthropic/claude-4-sonnet-20250522 (C)2025.1065491231253236
EvoPromptBackbone=google/gemini-2.5-flash (G)2025.1065481030242734
PromptbreederBackbone=google/gemini-2.5-flash (G)2025.1063471029233034
TextGradBackbone=anthropic/claude-4-sonnet-20250522 (C)2025.1062471330182533
TextGradBackbone=google/gemini-2.5-flash (G)2025.1061451128162331
ReflexionBackbone=anthropic/claude-4-sonnet-20250522 (C)2025.1060441127323134
ReflexionBackbone=google/gemini-2.5-flash (G)2025.105841925302932
GRPO (online)Backbone=qwen/qwen3-32b, Evaluation Protocol=Weight-update Online Fine-Tuning2025.105538722302630
SummaryBackbone=anthropic/claude-4-sonnet-20250522 (C)2025.1047351316272227
SummaryBackbone=google/gemini-2.5-flash (G)2025.1045331215252125
SFT (online)Backbone=qwen/qwen3-32b, Evaluation Protocol=Weight-update Online Fine-Tuning2025.1040301118221923
RAGBackbone=anthropic/claude-4-sonnet-20250522 (C)2025.103425810201619
RAGBackbone=google/gemini-2.5-flash (G)2025.10322479181518
MemoryBackbone=anthropic/claude-4-sonnet-20250522 (C)2025.10262057141114
MemoryBackbone=google/gemini-2.5-flash (G)2025.10251846131013
StaticBackbone=anthropic/claude-4-sonnet-20250522 (C)2025.1023164612912
StaticBackbone=google/gemini-2.5-flash (G)2025.1021153511811