List your company

Question Answering on WQ (test)

76.6AUROCSemantic Token Clustering
51.95258.35164.7571.149Mar 20, 2026

Evaluation results

MethodLinks
Semantic Token ClusteringLLM Backbone=Qwen2.5-7B2026.0376.6
CCPLLM Backbone=Qwen2.5-7B2026.0376.4
Semantic Token ClusteringLLM Backbone=Qwen2.5-3B2026.0376.2
EigVLLM Backbone=Qwen2.5-7B2026.0375.3
CCPLLM Backbone=Qwen2.5-3B2026.0375.2
EigVLLM Backbone=Qwen2.5-3B2026.0374.4
Semantic Token ClusteringLLM Backbone=Llama-3-8B2026.0374.4
Semantic Token ClusteringLLM Backbone=Qwen2.5-14B2026.0374.2
SARLLM Backbone=Qwen2.5-7B2026.0374
CCPLLM Backbone=Qwen2.5-14B2026.0374
EigVLLM Backbone=Qwen2.5-14B2026.0373.5
sentenceSARLLM Backbone=Qwen2.5-7B2026.0373.5
Semantic Token ClusteringLLM Backbone=Llama-2-7B2026.0373.5
DegLLM Backbone=Qwen2.5-7B2026.0373.4
Predictive EntropyLLM Backbone=Qwen2.5-7B2026.0373.1
SARLLM Backbone=Llama-3-8B2026.0373.1
LN EntropyLLM Backbone=Qwen2.5-7B2026.0372.9
Semantic EntropyLLM Backbone=Qwen2.5-3B2026.0372.8
EigVLLM Backbone=Llama-2-7B2026.0372.8
EigVLLM Backbone=Llama-3-8B2026.0372.8
EigVLLM Backbone=Mistral-7B2026.0372.8
CCPLLM Backbone=Llama-3-8B2026.0372.7
SARLLM Backbone=Qwen2.5-14B2026.0372.5
PerplexityLLM Backbone=Qwen2.5-7B2026.0372.4
Semantic EntropyLLM Backbone=Qwen2.5-7B2026.0372.4
PerplexityLLM Backbone=Qwen2.5-14B2026.0372.2
ConULLM Backbone=Qwen2.5-7B2026.0372.2
Semantic Token ClusteringLLM Backbone=Mistral-7B2026.0372.2
DegLLM Backbone=Qwen2.5-3B2026.0372.1
sentenceSARLLM Backbone=Llama-3-8B2026.0372.1
Semantic EntropyLLM Backbone=Qwen2.5-14B2026.0372
sentenceSARLLM Backbone=Qwen2.5-3B2026.0372
LN EntropyLLM Backbone=Llama-3-8B2026.0371.8
DegLLM Backbone=Llama-3-8B2026.0371.8
CCPLLM Backbone=Llama-2-7B2026.0371.6
Predictive EntropyLLM Backbone=Llama-3-8B2026.0371.5
Predictive EntropyLLM Backbone=Qwen2.5-3B2026.0371.5
P(True)LLM Backbone=Qwen2.5-7B2026.0371.5
CCPLLM Backbone=Mistral-7B2026.0371.5
LN EntropyLLM Backbone=Qwen2.5-14B2026.0371.4
DegLLM Backbone=Mistral-7B2026.0371.4
sentenceSARLLM Backbone=Qwen2.5-14B2026.0371.3
DegLLM Backbone=Llama-2-7B2026.0371.1
ConULLM Backbone=Qwen2.5-3B2026.0371
sentenceSARLLM Backbone=Llama-2-7B2026.0371
tokenSARLLM Backbone=Qwen2.5-14B2026.0370.9
Eigen ScoreLLM Backbone=Qwen2.5-7B2026.0370.9
Predictive EntropyLLM Backbone=Qwen2.5-14B2026.0370.8
DegLLM Backbone=Qwen2.5-14B2026.0370.8
P(True)LLM Backbone=Qwen2.5-14B2026.0370.7
Semantic EntropyLLM Backbone=Llama-2-7B2026.0370.5
SARLLM Backbone=Qwen2.5-3B2026.0370.5
Predictive EntropyLLM Backbone=Llama-2-7B2026.0370.4
ConULLM Backbone=Qwen2.5-14B2026.0370.3
Semantic EntropyLLM Backbone=Mistral-7B2026.0370.3
Eigen ScoreLLM Backbone=Qwen2.5-3B2026.0370.2
Semantic EntropyLLM Backbone=Llama-3-8B2026.0370.1
PerplexityLLM Backbone=Llama-3-8B2026.0369.9
ConULLM Backbone=Mistral-7B2026.0369.9
EccLLM Backbone=Qwen2.5-7B2026.0369.9
sentenceSARLLM Backbone=Mistral-7B2026.0369.7
SARLLM Backbone=Llama-2-7B2026.0369.5
LN EntropyLLM Backbone=Qwen2.5-3B2026.0369.4
tokenSARLLM Backbone=Qwen2.5-7B2026.0369.4
Eigen ScoreLLM Backbone=Llama-3-8B2026.0369.3
EccLLM Backbone=Llama-2-7B2026.0369.2
Predictive EntropyLLM Backbone=Mistral-7B2026.0369.1
EccLLM Backbone=Qwen2.5-3B2026.0369.1
Eigen ScoreLLM Backbone=Llama-2-7B2026.0369
EccLLM Backbone=Mistral-7B2026.0368.9
LN EntropyLLM Backbone=Llama-2-7B2026.0368.4
tokenSARLLM Backbone=Llama-3-8B2026.0368.4
Eigen ScoreLLM Backbone=Qwen2.5-14B2026.0368.4
SARLLM Backbone=Mistral-7B2026.0368.3
Eigen ScoreLLM Backbone=Mistral-7B2026.0368.2
ConULLM Backbone=Llama-2-7B2026.0368.1
EccLLM Backbone=Llama-3-8B2026.0368
ConULLM Backbone=Llama-3-8B2026.0367.8
PerplexityLLM Backbone=Qwen2.5-3B2026.0367.3
EccLLM Backbone=Qwen2.5-14B2026.0367.2
LN EntropyLLM Backbone=Mistral-7B2026.0366.5
PerplexityLLM Backbone=Llama-2-7B2026.0366.4
PerplexityLLM Backbone=Mistral-7B2026.0365.1
tokenSARLLM Backbone=Qwen2.5-3B2026.0364.3
tokenSARLLM Backbone=Llama-2-7B2026.0363.4
P(True)LLM Backbone=Qwen2.5-3B2026.0362.5
tokenSARLLM Backbone=Mistral-7B2026.0361.4
P(True)LLM Backbone=Mistral-7B2026.0356.2
P(True)LLM Backbone=Llama-3-8B2026.0356
P(True)LLM Backbone=Llama-2-7B2026.0352.9