Text Classification on N24News (test)
81.42Macro F1Voyage-4-nano + History-Aware Refinement (M3)
Evaluation Results
| Method | Links | |
|---|---|---|
| Voyage-4-nano + History-Aware Refinement (M3)Backbone=Voyage-4-nano, Refinement LLM=mistral, k=3, m=22026.04 | 81.42 | |
| KaLM-Embedding-v2.5 + History-Aware Refinement (M3)Backbone=KaLM-Embedding-v2.5, Refinement LLM=mistral, k=3, m=22026.04 | 81.18 | |
| KaLM-Embedding-v2.5 + Confusion-Aware Refinement (M2)Backbone=KaLM-Embedding-v2.5, Refinement LLM=mistral, k=22026.04 | 81.11 | |
| Voyage-4-nano + Example-Guided Refinement (M1)Backbone=Voyage-4-nano, Refinement LLM=gemma2026.04 | 80.9 | |
| Voyage-4-nano + Confusion-Aware Refinement (M2)Backbone=Voyage-4-nano, Refinement LLM=gemma, k=12026.04 | 80.29 | |
| KaLM-Embedding-v2.5 + Example-Guided Refinement (M1)Backbone=KaLM-Embedding-v2.5, Refinement LLM=qwen2026.04 | 80.06 | |
| Qwen3-Embedding-0.6B + History-Aware Refinement (M3)Backbone=Qwen3-Embedding-0.6B, Refinement LLM=mistral, k=5, m=42026.04 | 77.68 | |
| Qwen3-Embedding-0.6B + Confusion-Aware Refinement (M2)Backbone=Qwen3-Embedding-0.6B, Refinement LLM=gemma, k=52026.04 | 77.34 | |
| KaLM-Embedding-v2.5 (Baseline)Backbone=KaLM-Embedding-v2.5, Refinement=None2026.04 | 77.3 | |
| Modernbert-embed-large + History-Aware Refinement (M3)Backbone=Modernbert-embed-large, Refinement LLM=mistral, k=5, m=22026.04 | 77.01 | |
| Mxbai-embed-large-v1 + History-Aware Refinement (M3)Backbone=Mxbai-embed-large-v1, Refinement LLM=mistral, k=2, m=12026.04 | 76.57 | |
| Qwen3-Embedding-0.6B + Example-Guided Refinement (M1)Backbone=Qwen3-Embedding-0.6B, Refinement LLM=gemma2026.04 | 76.41 | |
| Modernbert-embed-large + Example-Guided Refinement (M1)Backbone=Modernbert-embed-large, Refinement LLM=gemma2026.04 | 76.28 | |
| Modernbert-embed-large + Confusion-Aware Refinement (M2)Backbone=Modernbert-embed-large, Refinement LLM=gemma, k=32026.04 | 75.94 | |
| Mxbai-embed-large-v1 + Confusion-Aware Refinement (M2)Backbone=Mxbai-embed-large-v1, Refinement LLM=mistral, k=32026.04 | 75.35 | |
| Mxbai-embed-large-v1 + Example-Guided Refinement (M1)Backbone=Mxbai-embed-large-v1, Refinement LLM=qwen2026.04 | 75.3 | |
| Voyage-4-nano (Baseline)Backbone=Voyage-4-nano, Refinement=None2026.04 | 75.24 | |
| Embeddinggemma-300m + History-Aware Refinement (M3)Backbone=Embeddinggemma-300m, Refinement LLM=qwen, k=4, m=52026.04 | 74.97 | |
| Embeddinggemma-300m + Example-Guided Refinement (M1)Backbone=Embeddinggemma-300m, Refinement LLM=glm-42026.04 | 74.35 | |
| Embeddinggemma-300m + Confusion-Aware Refinement (M2)Backbone=Embeddinggemma-300m, Refinement LLM=gemma, k=32026.04 | 73.72 | |
| Qwen3-Embedding-0.6B (Baseline)Backbone=Qwen3-Embedding-0.6B, Refinement=None2026.04 | 73.54 | |
| Modernbert-embed-large (Baseline)Backbone=Modernbert-embed-large, Refinement=None2026.04 | 72.36 | |
| Alibaba-NLP-gte + History-Aware Refinement (M3)Backbone=Alibaba-NLP-gte, Refinement LLM=gemma, k=2, m=22026.04 | 71.59 | |
| Embeddinggemma-300m (Baseline)Backbone=Embeddinggemma-300m, Refinement=None2026.04 | 71.09 | |
| Alibaba-NLP-gte + Confusion-Aware Refinement (M2)Backbone=Alibaba-NLP-gte, Refinement LLM=gemma, k=32026.04 | 70.47 | |
| Mxbai-embed-large-v1 (Baseline)Backbone=Mxbai-embed-large-v1, Refinement=None2026.04 | 69.76 | |
| Multilingual-e5-large + Confusion-Aware Refinement (M2)Backbone=Multilingual-e5-large, Refinement LLM=gemma, k=32026.04 | 69.65 | |
| Multilingual-e5-large + History-Aware Refinement (M3)Backbone=Multilingual-e5-large, Refinement LLM=glm-4, k=3, m=12026.04 | 69.2 | |
| Snowflake-arctic-embed-v2.0 + History-Aware Refinement (M3)Backbone=Snowflake-arctic-embed-v2.0, Refinement LLM=mistral, k=5, m=12026.04 | 68.19 | |
| BAAI bge-m3 + History-Aware Refinement (M3)Backbone=BAAI bge-m3, Refinement LLM=mistral, k=4, m=22026.04 | 67.58 | |
| Granite-embedding-278m + Confusion-Aware Refinement (M2)Backbone=Granite-embedding-278m, Refinement LLM=glm-4, k=42026.04 | 67.53 | |
| Alibaba-NLP-gte + Example-Guided Refinement (M1)Backbone=Alibaba-NLP-gte, Refinement LLM=gemma2026.04 | 67.19 | |
| Solon-embeddings-large-0.1 + History-Aware Refinement (M3)Backbone=Solon-embeddings-large-0.1, Refinement LLM=gemma, k=3, m=22026.04 | 66.9 | |
| Granite-embedding-278m + History-Aware Refinement (M3)Backbone=Granite-embedding-278m, Refinement LLM=glm-4, k=3, m=12026.04 | 66.68 | |
| Multilingual-e5-large + Example-Guided Refinement (M1)Backbone=Multilingual-e5-large, Refinement LLM=qwen2026.04 | 66.62 | |
| BAAI bge-m3 + Example-Guided Refinement (M1)Backbone=BAAI bge-m3, Refinement LLM=glm-42026.04 | 66.57 | |
| Snowflake-arctic-embed-v2.0 + Confusion-Aware Refinement (M2)Backbone=Snowflake-arctic-embed-v2.0, Refinement LLM=mistral, k=32026.04 | 66.44 | |
| BAAI bge-m3 + Confusion-Aware Refinement (M2)Backbone=BAAI bge-m3, Refinement LLM=glm-4, k=42026.04 | 66.06 | |
| Snowflake-arctic-embed-v2.0 + Example-Guided Refinement (M1)Backbone=Snowflake-arctic-embed-v2.0, Refinement LLM=qwen2026.04 | 65.22 | |
| Solon-embeddings-large-0.1 + Confusion-Aware Refinement (M2)Backbone=Solon-embeddings-large-0.1, Refinement LLM=qwen, k=32026.04 | 64.06 | |
| Granite-embedding-278m + Example-Guided Refinement (M1)Backbone=Granite-embedding-278m, Refinement LLM=qwen2026.04 | 63.69 | |
| Llama-nemotron-embed-1b-v2 + History-Aware Refinement (M3)Backbone=Llama-nemotron-embed-1b-v2, Refinement LLM=gemma, k=3, m=22026.04 | 63.17 | |
| Solon-embeddings-large-0.1 + Example-Guided Refinement (M1)Backbone=Solon-embeddings-large-0.1, Refinement LLM=qwen2026.04 | 62.56 | |
| Llama-nemotron-embed-1b-v2 + Confusion-Aware Refinement (M2)Backbone=Llama-nemotron-embed-1b-v2, Refinement LLM=glm-4, k=42026.04 | 61.52 | |
| Llama-nemotron-embed-1b-v2 + Example-Guided Refinement (M1)Backbone=Llama-nemotron-embed-1b-v2, Refinement LLM=gemma2026.04 | 60.52 | |
| BAAI bge-m3 (Baseline)Backbone=BAAI bge-m3, Refinement=None2026.04 | 58.51 | |
| Snowflake-arctic-embed-v2.0 (Baseline)Backbone=Snowflake-arctic-embed-v2.0, Refinement=None2026.04 | 57.21 | |
| Solon-embeddings-large-0.1 (Baseline)Backbone=Solon-embeddings-large-0.1, Refinement=None2026.04 | 55.99 | |
| Granite-embedding-278m (Baseline)Backbone=Granite-embedding-278m, Refinement=None2026.04 | 55.38 | |
| Alibaba-NLP-gte (Baseline)Backbone=Alibaba-NLP-gte, Refinement=None2026.04 | 54.47 | |
| Multilingual-e5-large (Baseline)Backbone=Multilingual-e5-large, Refinement=None2026.04 | 49.3 | |
| Llama-nemotron-embed-1b-v2 (Baseline)Backbone=Llama-nemotron-embed-1b-v2, Refinement=None2026.04 | 47.39 |