Text Classification on MultiWD
60.35AccuracyCalibration w/ ADAMAB
Evaluation Results
| Method | Links | |
|---|---|---|
| Calibration w/ ADAMABProtocol=Few-shot, Backbone=QWen-3-emb-06b, # Params=+1.180M2026.02 | 60.35 | |
| Calibration w/ ADAMABProtocol=Few-shot, Backbone=OpenAI-text-embedding-3-small, # Params=+2.654M2026.02 | 59.91 | |
| Calibration w/ ADAMABProtocol=Zero-shot, Backbone=OpenAI-text-embedding-3-small, # Params=+2.654M2026.02 | 58.15 | |
| Calibration w/ ADAMABProtocol=Zero-shot, Backbone=QWen-3-emb-06b, # Params=+1.180M2026.02 | 58.15 | |
| Calibration w/ random aug.Protocol=Zero-shot, Backbone=OpenAI-text-embedding-3-small, # Params=+2.654M2026.02 | 56.83 | |
| Calibration w/ random aug.Protocol=Few-shot, Backbone=OpenAI-text-embedding-3-small, # Params=+2.654M2026.02 | 55.51 | |
| Calibration w/ random aug.Protocol=Few-shot, Backbone=QWen-3-emb-06b, # Params=+1.180M2026.02 | 55.07 | |
| GPT-4o-miniProtocol=Few-shot, # Params=n/a2026.02 | 51.54 | |
| Calibration w/ only init setProtocol=Zero-shot, Backbone=OpenAI-text-embedding-3-small, # Params=+2.654M2026.02 | 50.66 | |
| Calibration w/ only init setProtocol=Few-shot, Backbone=OpenAI-text-embedding-3-small, # Params=+2.654M2026.02 | 50.22 | |
| QWen-3-emb-06bProtocol=Few-shot, # Params=0.596B2026.02 | 49.34 | |
| Calibration w/ random aug.Protocol=Zero-shot, Backbone=QWen-3-emb-06b, # Params=+1.180M2026.02 | 47.58 | |
| Gemini2.0-Flash-LiteProtocol=Few-shot, # Params=n/a2026.02 | 46.39 | |
| Mistral-SmallProtocol=Few-shot, # Params=24B2026.02 | 45.82 | |
| Calibration w/ only init setProtocol=Zero-shot, Backbone=QWen-3-emb-06b, # Params=+1.180M2026.02 | 44.93 | |
| Calibration w/ only init setProtocol=Few-shot, Backbone=QWen-3-emb-06b, # Params=+1.180M2026.02 | 44.93 | |
| Gemini2.0-Flash-LiteProtocol=Zero-shot, # Params=n/a2026.02 | 39.65 | |
| OpenAI-text-embedding-3-smallProtocol=Zero-shot, # Params=n/a2026.02 | 39.21 | |
| GPT-4o-miniProtocol=Zero-shot, # Params=n/a2026.02 | 37.89 | |
| OpenAI-text-embedding-3-smallProtocol=Few-shot, # Params=n/a2026.02 | 37.88 | |
| Llama-3.2Protocol=Few-shot, # Params=90B2026.02 | 37 | |
| QWen-3-emb-06bProtocol=Zero-shot, # Params=0.596B2026.02 | 36.56 | |
| Mistral-SmallProtocol=Zero-shot, # Params=24B2026.02 | 30.4 | |
| Llama-3.2Protocol=Zero-shot, # Params=90B2026.02 | 28.19 | |
| Cohere-v3.5Protocol=Few-shot, # Params=n/a2026.02 | 25.55 | |
| Cohere-v3.5Protocol=Zero-shot, # Params=n/a2026.02 | 3.08 | |
| BGE-reranker-v2-m3Protocol=Zero-shot, # Params=0.568B2026.02 | 3.08 | |
| BGE-reranker-v2-m3Protocol=Few-shot, # Params=0.568B2026.02 | 3.08 |