Temporal Reasoning (date) on Synthetic Temporal Dataset (10-fold CV) benchmark leaderboard · SOTA2 Research

Temporal Reasoning (date) on Synthetic Temporal Dataset (10-fold CV)

1.103Negative Log ScoreLlama-3.2-3B

Evaluation Results

Method	Links
Llama-3.2-3B2025.10		1.103	1
gemma-2-2b2025.10		1.423	1
Qwen2.5-3B2025.10		1.764	8
Qwen2.5-3B-IT2025.10		1.964	32
gemma-2-2b-IT2025.10		2.036	58
Llama-3.2-3B2025.10		2.116	39
Llama-3.2-3B-IT2025.10		2.192	49
Llama-3.1-8B-IT2025.10		2.27	31
Llama-3.1-8B-IT2025.10		2.483	39
Llama-3.2-3B-IT2025.10		2.502	81
Llama-3.1-70B-IT2025.10		2.636	18
Qwen2.5-3B2025.10		2.797	21
Llama-3.1-70B-IT2025.10		2.909	93
gemma-2-2b-IT2025.10		2.969	38
gemma-2-2b2025.10		3.142	31
Qwen2.5-3B-IT2025.10		3.267	36
Llama-3.1-8B-IT2025.10		6.018	49
Llama-3.1-70B-IT2025.10		6.219	82
Llama-3.2-3B-IT2025.10		6.525	51
Llama-3.2-3B2025.10		6.702	37
gemma-2-2b2025.10		6.948	33
gemma-2-2b-IT2025.10		7.207	36
Qwen2.5-3B2025.10		7.254	23
Qwen2.5-3B-IT2025.10		7.548	25