List your company

Causal Reasoning on Cladder (Accuracy)

82.7AccuracyCDCR
42.45252.90163.3573.799May 5, 2026May 8, 2026May 11, 2026May 14, 2026May 17, 2026May 20, 2026May 24, 2026

Evaluation results

MethodLinks
CDCRBase Model=Qwen3-8B, Data=CDCR2026.0582.7
Graph-Guided (Ours)2026.0582.3
Qwen3-32BBase Model=Qwen3-32B2026.0582.3
CDCRBase Model=Qwen3-4B, Data=CDCR2026.0582
CDCRBase Model=Olmo3-7B-Instruct, Data=CDCR2026.0581.7
UNICOBase Model=Qwen3-8B, Data=UNICO2026.0581
UNICOBase Model=Qwen3-4B, Data=UNICO2026.0579
Olmo3.1-32B-InstructBase Model=Olmo3.1-32B-Instruct2026.0578
UNICOBase Model=Olmo3-7B-Instruct, Data=UNICO2026.0577.3
CauGymBase Model=Qwen3-8B, Data=CauGym2026.0576.7
CauGymBase Model=Qwen3-4B, Data=CauGym2026.0575.2
OriginalBase Model=Qwen3-8B, Data=Original2026.0574.9
CauGymBase Model=Olmo3-7B-Instruct, Data=CauGym2026.0574.3
OriginalBase Model=Olmo3-7B-Instruct, Data=Original2026.0574.2
OriginalBase Model=Qwen3-4B, Data=Original2026.0571.2
Causal CoT2026.0570.4
GPT-42026.0562
GPT-3.52026.0552.2
Alpaca2026.0544.7
LLaMa2026.0544