Variant Classification on COHA Historical American English (Pre-hoc)
98Accuracy (Sample 1)LLM-assisted corpus annotation pipeline
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| LLM-assisted corpus annotation pipelineModel=GPT-5, Evaluation=Pre-hoc, Sample Size (N)=100, Averaging=Macro-averaged2025.10 | 98 | 99 | 99 | 98.67 | 97.2 | 98.12 | 97.5 | 0.9676 |