RST Discourse Parsing on RST-DT (test)
79.8Span F1Llama 2 (70B)
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| Llama 2 (70B)Parsing Strategy=Bottom-up, Backbone=Llama 2, Parameters=70B, Quantization=QLoRA 4-bit2024.03 | 79.8 | 70.4 | 60 | 58.1 | |
| Llama 2 (70B)Parsing Strategy=Top-down, Backbone=Llama 2, Parameters=70B, Quantization=QLoRA 4-bit2024.03 | 78.8 | 68.7 | 57.7 | 56 | |
| Llama 2 (13B)Parsing Strategy=Top-down, Backbone=Llama 2, Parameters=13B, Quantization=QLoRA 4-bit2024.03 | 78.6 | 67.9 | 57.7 | 55.6 | |
| Kobayashi et al.Parsing Strategy=Top-down, Backbone=DeBERTa-base, Parameters=140M2024.03 | 78.5 | 67.9 | 56.6 | 54.4 | |
| Llama 2 (13B)Parsing Strategy=Bottom-up, Backbone=Llama 2, Parameters=13B, Quantization=QLoRA 4-bit2024.03 | 78.3 | 68.1 | 57.8 | 56 | |
| Llama 2 (7B)Parsing Strategy=Bottom-up, Backbone=Llama 2, Parameters=7B, Quantization=QLoRA 4-bit2024.03 | 78.2 | 67.5 | 57.6 | 55.8 | |
| Kobayashi et al.Parsing Strategy=Bottom-up, Backbone=DeBERTa-base, Parameters=140M2024.03 | 77.8 | 68 | 57.3 | 55.4 | |
| Liu et al.Parsing Strategy=Top-down, Backbone=XLM-ROBERTa-base, Parameters=125M2024.03 | 76.5 | 65.2 | 54.2 | — | |
| Guz et al.Parsing Strategy=Bottom-up, Backbone=SpanBERT-base, Parameters=110M2024.03 | 76.5 | 65.9 | 54.8 | — | |
| Yu et al.Parsing Strategy=Bottom-up, Backbone=XLNet-base, Parameters=110M2024.03 | 76.4 | 66.1 | 54.5 | 53.5 | |
| Llama 2 (7B)Parsing Strategy=Top-down, Backbone=Llama 2, Parameters=7B, Quantization=QLoRA 4-bit2024.03 | 76.3 | 65.4 | 55.2 | 53.4 | |
| Yu et al.Parsing Strategy=Top-down, Backbone=XLNet-base, Parameters=110M2024.03 | 72.9 | 62.7 | 52.5 | 50.5 |