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

Backward Transition Handling on PFDial-H (test)

76.25Accuracy

Qwen2.5-7B

Updated 1mo ago

Evaluation Results

Method	Links
Qwen2.5-7B 2025.03		76.25	86.36	72.41
Qwen2.5-3B 2025.03		75	90.91	68.97
Qwen2.5-1.5B 2025.03		70	72.73	68.97
Qwen2.5-7B 2025.03		66.25	90.91	56.9
Qwen2.5-0.5B 2025.03		65	72.73	62.07
GPT-4o 2025.03		58.75	68.18	55.17
Qwen2.5-0.5B 2025.03		58.75	77.27	51.72
Claude-3.5-Sonnet 2025.03		57.5	59.09	56.9
Qwen2.5-3B 2025.03		57.5	59.09	56.9
GPT-3.5-Turbo 2025.03		51.25	54.55	50
Qwen2.5-1.5B 2025.03		47.5	45.45	48.28
DeepSeek-v3 2025.03		40	31.82	43.1
Qwen2.5-7B-Instruct 2025.03		31.25	31.82	31.03
Gemini-2.0-Flash-Exp 2025.03		26.25	22.73	27.59
LLaMA-3.1-8B-Instruct 2025.03		15	13.64	15.52