List your company

Question Answering on TQA (test)

90.2AUROCSemantic Token Clustering
51.82461.78771.7581.713Mar 20, 2026

Evaluation results

MethodLinks
Semantic Token ClusteringLLM Backbone=Qwen2.5-7B2026.0390.2
Semantic Token ClusteringLLM Backbone=Qwen2.5-14B2026.0390
CCPLLM Backbone=Qwen2.5-7B2026.0389.5
Semantic Token ClusteringLLM Backbone=Mistral-7B2026.0389.3
CCPLLM Backbone=Mistral-7B2026.0389.2
PerplexityLLM Backbone=Qwen2.5-7B2026.0388.8
CCPLLM Backbone=Qwen2.5-14B2026.0388.8
SARLLM Backbone=Qwen2.5-7B2026.0388.4
LN EntropyLLM Backbone=Qwen2.5-7B2026.0388.3
sentenceSARLLM Backbone=Qwen2.5-7B2026.0388.3
Semantic Token ClusteringLLM Backbone=Llama-3-8B2026.0388.2
PerplexityLLM Backbone=Qwen2.5-14B2026.0388.1
SARLLM Backbone=Mistral-7B2026.0388.1
Predictive EntropyLLM Backbone=Qwen2.5-7B2026.0388
SARLLM Backbone=Qwen2.5-14B2026.0387.9
CCPLLM Backbone=Llama-3-8B2026.0387.8
Semantic Token ClusteringLLM Backbone=Llama-2-7B2026.0387.8
sentenceSARLLM Backbone=Qwen2.5-14B2026.0387.7
Semantic Token ClusteringLLM Backbone=Qwen2.5-3B2026.0387.6
LN EntropyLLM Backbone=Qwen2.5-14B2026.0387.5
CCPLLM Backbone=Llama-2-7B2026.0387.4
Predictive EntropyLLM Backbone=Qwen2.5-14B2026.0387.3
PerplexityLLM Backbone=Mistral-7B2026.0387.2
SARLLM Backbone=Llama-3-8B2026.0387.2
tokenSARLLM Backbone=Qwen2.5-7B2026.0387.1
LN EntropyLLM Backbone=Mistral-7B2026.0387
P(True)LLM Backbone=Qwen2.5-14B2026.0386.9
sentenceSARLLM Backbone=Mistral-7B2026.0386.9
CCPLLM Backbone=Qwen2.5-3B2026.0386.9
PerplexityLLM Backbone=Llama-2-7B2026.0386.8
tokenSARLLM Backbone=Qwen2.5-14B2026.0386.7
SARLLM Backbone=Llama-2-7B2026.0386.7
PerplexityLLM Backbone=Llama-3-8B2026.0386.5
Predictive EntropyLLM Backbone=Mistral-7B2026.0386.4
LN EntropyLLM Backbone=Llama-2-7B2026.0386.2
LN EntropyLLM Backbone=Llama-3-8B2026.0386.2
tokenSARLLM Backbone=Mistral-7B2026.0386
sentenceSARLLM Backbone=Llama-2-7B2026.0385.9
sentenceSARLLM Backbone=Llama-3-8B2026.0385.9
PerplexityLLM Backbone=Qwen2.5-3B2026.0385.6
Predictive EntropyLLM Backbone=Llama-2-7B2026.0385.5
Predictive EntropyLLM Backbone=Llama-3-8B2026.0385.4
sentenceSARLLM Backbone=Qwen2.5-3B2026.0385.4
SARLLM Backbone=Qwen2.5-3B2026.0385.4
LN EntropyLLM Backbone=Qwen2.5-3B2026.0385.3
tokenSARLLM Backbone=Llama-2-7B2026.0385.3
tokenSARLLM Backbone=Llama-3-8B2026.0385.3
Predictive EntropyLLM Backbone=Qwen2.5-3B2026.0385
EigVLLM Backbone=Qwen2.5-7B2026.0384.9
EigVLLM Backbone=Mistral-7B2026.0384.5
Semantic EntropyLLM Backbone=Qwen2.5-7B2026.0384.4
Semantic EntropyLLM Backbone=Mistral-7B2026.0383.9
DegLLM Backbone=Qwen2.5-7B2026.0383.9
EigVLLM Backbone=Qwen2.5-3B2026.0383.7
DegLLM Backbone=Mistral-7B2026.0383.7
EigVLLM Backbone=Qwen2.5-14B2026.0383.5
P(True)LLM Backbone=Qwen2.5-7B2026.0383.2
Semantic EntropyLLM Backbone=Qwen2.5-3B2026.0383.2
Semantic EntropyLLM Backbone=Qwen2.5-14B2026.0383.2
EigVLLM Backbone=Llama-2-7B2026.0383.2
Eigen ScoreLLM Backbone=Qwen2.5-7B2026.0383.2
tokenSARLLM Backbone=Qwen2.5-3B2026.0383.1
DegLLM Backbone=Qwen2.5-3B2026.0383.1
DegLLM Backbone=Qwen2.5-14B2026.0382.7
Semantic EntropyLLM Backbone=Llama-2-7B2026.0382.6
EigVLLM Backbone=Llama-3-8B2026.0382.5
DegLLM Backbone=Llama-2-7B2026.0382.4
Semantic EntropyLLM Backbone=Llama-3-8B2026.0382.3
Eigen ScoreLLM Backbone=Qwen2.5-3B2026.0382
DegLLM Backbone=Llama-3-8B2026.0381.8
EccLLM Backbone=Qwen2.5-7B2026.0381.5
Eigen ScoreLLM Backbone=Qwen2.5-14B2026.0381.5
EccLLM Backbone=Mistral-7B2026.0381
EccLLM Backbone=Qwen2.5-3B2026.0380.9
Eigen ScoreLLM Backbone=Mistral-7B2026.0380.9
Eigen ScoreLLM Backbone=Llama-2-7B2026.0380.7
Eigen ScoreLLM Backbone=Llama-3-8B2026.0380.5
EccLLM Backbone=Qwen2.5-14B2026.0380.2
EccLLM Backbone=Llama-2-7B2026.0379.9
EccLLM Backbone=Llama-3-8B2026.0379.2
ConULLM Backbone=Qwen2.5-7B2026.0377.1
ConULLM Backbone=Mistral-7B2026.0375.8
ConULLM Backbone=Qwen2.5-14B2026.0375.4
ConULLM Backbone=Llama-3-8B2026.0375.3
ConULLM Backbone=Llama-2-7B2026.0375
ConULLM Backbone=Qwen2.5-3B2026.0374.9
P(True)LLM Backbone=Qwen2.5-3B2026.0373.9
P(True)LLM Backbone=Llama-3-8B2026.0364
P(True)LLM Backbone=Mistral-7B2026.0358.7
P(True)LLM Backbone=Llama-2-7B2026.0353.3