Share your thoughts, 1 month free Claude Pro on usSee more

Rule-discovery on Blicket baseline (test)

86Task Success

Qwen3-32B

Updated 3mo ago

Evaluation Results

Method	Links
Qwen3-32B 2026.04		86	12.98	1.31
Qwen3-32B + 32B TiO 2026.04		77	11.02	0.88
Qwen3-32B 2026.04		57	12.18	0.54
Qwen3-8B 2026.04		40	10.73	0.57
Qwen3-8B + 32B TiO 2026.04		38	11.84	0.51
Qwen3-8B + 32B 2026.04		36	10.25	0.53
Qwen3-8B + 8B TiO 2026.04		34	13.96	0.48
Qwen3-8B 2026.04		32	10.26	0.44