Story-level evaluation on HANNA benchmark leaderboard · SOTA2 Research

Story-level evaluation on HANNA

0.678Coherence (RP)BATCHEVAL

Evaluation Results

Method	Links
BATCHEVAL2023.12		0.678	0.625	0.702	0.679	0.546	0.543	0.368	0.381	0.617	0.605	0.625	0.575	0.589	0.568	0.0538
Closer Look2023.12		0.595	0.591	0.579	0.597	0.498	0.478	0.28	0.339	0.605	0.607	0.619	0.568	0.529	0.53	0.0835
G-Eval2023.12		0.572	0.578	0.582	0.584	0.453	0.461	0.311	0.347	0.562	0.591	0.602	0.557	0.514	0.52	0.0772
BERTScore2023.12		0.358	0.293	0.201	0.188	0.308	0.303	0.302	0.29	0.308	0.331	0.501	0.472	0.33	0.313	—
METEOR2023.12		0.335	0.273	0.202	0.19	0.304	0.282	0.285	0.283	0.316	0.338	0.52	0.482	0.307	0.307	—
BLEU-42023.12		0.22	0.218	0.135	0.175	0.242	0.216	0.178	0.224	0.242	0.27	0.362	0.273	0.23	0.229	—