Sentiment Classification on TwSenti
61.32F1 ScoreK-Means + Feedback-ICL
Evaluation results
| Method | Links | |
|---|---|---|
| K-Means + Feedback-ICLBackbone=Llama-2 13B Chat, Evaluation Protocol=In-context learning, Number of in-context examples=4, Retrieval Strategy=K-Means2024.06 | 61.32 | |
| Random + Feedback-ICLBackbone=Llama-2 13B Chat, Evaluation Protocol=In-context learning, Number of in-context examples=4, Retrieval Strategy=Random2024.06 | 60.33 | |
| BM25 + Feedback-ICLBackbone=Llama-2 13B Chat, Evaluation Protocol=In-context learning, Number of in-context examples=4, Retrieval Strategy=BM252024.06 | 59.2 | |
| MMR + Feedback-ICLBackbone=Llama-2 13B Chat, Evaluation Protocol=In-context learning, Number of in-context examples=4, Retrieval Strategy=MMR2024.06 | 56.84 | |
| K-MeansBackbone=Llama-2 13B Chat, Evaluation Protocol=In-context learning, Number of in-context examples=4, Retrieval Strategy=K-Means2024.06 | 56.26 | |
| BM25Backbone=Llama-2 13B Chat, Evaluation Protocol=In-context learning, Number of in-context examples=4, Retrieval Strategy=BM252024.06 | 55.35 | |
| RandomBackbone=Llama-2 13B Chat, Evaluation Protocol=In-context learning, Number of in-context examples=4, Retrieval Strategy=Random2024.06 | 55.27 | |
| SBERT + Feedback-ICLBackbone=Llama-2 13B Chat, Evaluation Protocol=In-context learning, Number of in-context examples=4, Retrieval Strategy=SBERT2024.06 | 55.08 | |
| BERT-FTBackbone=BERT-base, Evaluation Protocol=Fine-tuning2024.06 | 54.54 | |
| MMRBackbone=Llama-2 13B Chat, Evaluation Protocol=In-context learning, Number of in-context examples=4, Retrieval Strategy=MMR2024.06 | 50.8 | |
| SBERTBackbone=Llama-2 13B Chat, Evaluation Protocol=In-context learning, Number of in-context examples=4, Retrieval Strategy=SBERT2024.06 | 50.13 |