Question Answering on WQ (test)
76.6AUROCSemantic Token Clustering
Evaluation Results
| Method | Links | |
|---|---|---|
| Semantic Token ClusteringLLM Backbone=Qwen2.5-7B2026.03 | 76.6 | |
| CCPLLM Backbone=Qwen2.5-7B2026.03 | 76.4 | |
| Semantic Token ClusteringLLM Backbone=Qwen2.5-3B2026.03 | 76.2 | |
| EigVLLM Backbone=Qwen2.5-7B2026.03 | 75.3 | |
| CCPLLM Backbone=Qwen2.5-3B2026.03 | 75.2 | |
| EigVLLM Backbone=Qwen2.5-3B2026.03 | 74.4 | |
| Semantic Token ClusteringLLM Backbone=Llama-3-8B2026.03 | 74.4 | |
| Semantic Token ClusteringLLM Backbone=Qwen2.5-14B2026.03 | 74.2 | |
| SARLLM Backbone=Qwen2.5-7B2026.03 | 74 | |
| CCPLLM Backbone=Qwen2.5-14B2026.03 | 74 | |
| EigVLLM Backbone=Qwen2.5-14B2026.03 | 73.5 | |
| sentenceSARLLM Backbone=Qwen2.5-7B2026.03 | 73.5 | |
| Semantic Token ClusteringLLM Backbone=Llama-2-7B2026.03 | 73.5 | |
| DegLLM Backbone=Qwen2.5-7B2026.03 | 73.4 | |
| Predictive EntropyLLM Backbone=Qwen2.5-7B2026.03 | 73.1 | |
| SARLLM Backbone=Llama-3-8B2026.03 | 73.1 | |
| LN EntropyLLM Backbone=Qwen2.5-7B2026.03 | 72.9 | |
| Semantic EntropyLLM Backbone=Qwen2.5-3B2026.03 | 72.8 | |
| EigVLLM Backbone=Llama-2-7B2026.03 | 72.8 | |
| EigVLLM Backbone=Llama-3-8B2026.03 | 72.8 | |
| EigVLLM Backbone=Mistral-7B2026.03 | 72.8 | |
| CCPLLM Backbone=Llama-3-8B2026.03 | 72.7 | |
| SARLLM Backbone=Qwen2.5-14B2026.03 | 72.5 | |
| PerplexityLLM Backbone=Qwen2.5-7B2026.03 | 72.4 | |
| Semantic EntropyLLM Backbone=Qwen2.5-7B2026.03 | 72.4 | |
| PerplexityLLM Backbone=Qwen2.5-14B2026.03 | 72.2 | |
| ConULLM Backbone=Qwen2.5-7B2026.03 | 72.2 | |
| Semantic Token ClusteringLLM Backbone=Mistral-7B2026.03 | 72.2 | |
| DegLLM Backbone=Qwen2.5-3B2026.03 | 72.1 | |
| sentenceSARLLM Backbone=Llama-3-8B2026.03 | 72.1 | |
| Semantic EntropyLLM Backbone=Qwen2.5-14B2026.03 | 72 | |
| sentenceSARLLM Backbone=Qwen2.5-3B2026.03 | 72 | |
| LN EntropyLLM Backbone=Llama-3-8B2026.03 | 71.8 | |
| DegLLM Backbone=Llama-3-8B2026.03 | 71.8 | |
| CCPLLM Backbone=Llama-2-7B2026.03 | 71.6 | |
| Predictive EntropyLLM Backbone=Llama-3-8B2026.03 | 71.5 | |
| Predictive EntropyLLM Backbone=Qwen2.5-3B2026.03 | 71.5 | |
| P(True)LLM Backbone=Qwen2.5-7B2026.03 | 71.5 | |
| CCPLLM Backbone=Mistral-7B2026.03 | 71.5 | |
| LN EntropyLLM Backbone=Qwen2.5-14B2026.03 | 71.4 | |
| DegLLM Backbone=Mistral-7B2026.03 | 71.4 | |
| sentenceSARLLM Backbone=Qwen2.5-14B2026.03 | 71.3 | |
| DegLLM Backbone=Llama-2-7B2026.03 | 71.1 | |
| ConULLM Backbone=Qwen2.5-3B2026.03 | 71 | |
| sentenceSARLLM Backbone=Llama-2-7B2026.03 | 71 | |
| tokenSARLLM Backbone=Qwen2.5-14B2026.03 | 70.9 | |
| Eigen ScoreLLM Backbone=Qwen2.5-7B2026.03 | 70.9 | |
| Predictive EntropyLLM Backbone=Qwen2.5-14B2026.03 | 70.8 | |
| DegLLM Backbone=Qwen2.5-14B2026.03 | 70.8 | |
| P(True)LLM Backbone=Qwen2.5-14B2026.03 | 70.7 | |
| Semantic EntropyLLM Backbone=Llama-2-7B2026.03 | 70.5 | |
| SARLLM Backbone=Qwen2.5-3B2026.03 | 70.5 | |
| Predictive EntropyLLM Backbone=Llama-2-7B2026.03 | 70.4 | |
| ConULLM Backbone=Qwen2.5-14B2026.03 | 70.3 | |
| Semantic EntropyLLM Backbone=Mistral-7B2026.03 | 70.3 | |
| Eigen ScoreLLM Backbone=Qwen2.5-3B2026.03 | 70.2 | |
| Semantic EntropyLLM Backbone=Llama-3-8B2026.03 | 70.1 | |
| PerplexityLLM Backbone=Llama-3-8B2026.03 | 69.9 | |
| ConULLM Backbone=Mistral-7B2026.03 | 69.9 | |
| EccLLM Backbone=Qwen2.5-7B2026.03 | 69.9 | |
| sentenceSARLLM Backbone=Mistral-7B2026.03 | 69.7 | |
| SARLLM Backbone=Llama-2-7B2026.03 | 69.5 | |
| LN EntropyLLM Backbone=Qwen2.5-3B2026.03 | 69.4 | |
| tokenSARLLM Backbone=Qwen2.5-7B2026.03 | 69.4 | |
| Eigen ScoreLLM Backbone=Llama-3-8B2026.03 | 69.3 | |
| EccLLM Backbone=Llama-2-7B2026.03 | 69.2 | |
| Predictive EntropyLLM Backbone=Mistral-7B2026.03 | 69.1 | |
| EccLLM Backbone=Qwen2.5-3B2026.03 | 69.1 | |
| Eigen ScoreLLM Backbone=Llama-2-7B2026.03 | 69 | |
| EccLLM Backbone=Mistral-7B2026.03 | 68.9 | |
| LN EntropyLLM Backbone=Llama-2-7B2026.03 | 68.4 | |
| tokenSARLLM Backbone=Llama-3-8B2026.03 | 68.4 | |
| Eigen ScoreLLM Backbone=Qwen2.5-14B2026.03 | 68.4 | |
| SARLLM Backbone=Mistral-7B2026.03 | 68.3 | |
| Eigen ScoreLLM Backbone=Mistral-7B2026.03 | 68.2 | |
| ConULLM Backbone=Llama-2-7B2026.03 | 68.1 | |
| EccLLM Backbone=Llama-3-8B2026.03 | 68 | |
| ConULLM Backbone=Llama-3-8B2026.03 | 67.8 | |
| PerplexityLLM Backbone=Qwen2.5-3B2026.03 | 67.3 | |
| EccLLM Backbone=Qwen2.5-14B2026.03 | 67.2 | |
| LN EntropyLLM Backbone=Mistral-7B2026.03 | 66.5 | |
| PerplexityLLM Backbone=Llama-2-7B2026.03 | 66.4 | |
| PerplexityLLM Backbone=Mistral-7B2026.03 | 65.1 | |
| tokenSARLLM Backbone=Qwen2.5-3B2026.03 | 64.3 | |
| tokenSARLLM Backbone=Llama-2-7B2026.03 | 63.4 | |
| P(True)LLM Backbone=Qwen2.5-3B2026.03 | 62.5 | |
| tokenSARLLM Backbone=Mistral-7B2026.03 | 61.4 | |
| P(True)LLM Backbone=Mistral-7B2026.03 | 56.2 | |
| P(True)LLM Backbone=Llama-3-8B2026.03 | 56 | |
| P(True)LLM Backbone=Llama-2-7B2026.03 | 52.9 |