ktlKTL / Eval-GCSC

Eval-GCSC: A New Metric Approach for Evaluating ChatGPT's Performance in Chinese Spelling Correction

Eval-GCSC

Eval-GCSC: A New Metric Approach for Evaluating ChatGPT's Performance in Chinese Spelling Correction

Data

Experiments were conducted on the widely used benchmark dataset SIGHAN13~15.

Eval-GCSC

We propose Eval-GCSC, a novel Evaluation metric for Generative error correction models, for the CSC.

Code

The code was subsequently released.