Quantifying and Auditing LLM Evaluation via Positive--Unlabeled Learning
Zilong Zhang, Yi-Ting Hung, Lei Ding, Chi-Kuang Yeh
stat.ML, cs.LG, stat.CO, stat.ME
Submitted: 2026-06-17
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Llemma: An Open Language Model For Mathematics
- Evaluation of Text Generation: A Survey
- M3-Embedding: Multi-Linguality, Multi-Functionality, Multi-Granularity Text Embeddings Through Self-Knowledge Distillation
- LM vs LM: Detecting Factual Errors via Cross Examination
- A Survey on LLM-as-a-Judge
- Impact of Positional Encoding: Clean and Adversarial Rademacher Complexity for Transformers under In-Context Regression
- Mistral 7B
- Prometheus 2: An Open Source Language Model Specialized in Evaluating Other Language Models
- Holistic Evaluation of Language Models
- G-Eval: NLG Evaluation using GPT-4 with Better Human Alignment
- Beyond Accuracy: Behavioral Testing of NLP models with CheckList
- Verbosity Bias in Preference Labeling by Large Language Models
- A Long Way to Go: Investigating Length Correlations in RLHF
- EmbeddingGemma: Powerful and Lightweight Text Representations
- Replacing Judges with Juries: Evaluating LLM Generations with a Panel of Diverse Models
- Qwen2.5 Technical Report
- BERTScore: Evaluating Text Generation with BERT
- Qwen3 Embedding: Advancing Text Embedding and Reranking Through Foundation Models
- JudgeLM: Fine-tuned Large Language Models are Scalable Judges
Related papers
- Behavior of prediction performance metrics with rare events
- Optimal Estimation of Generic Dynamics by Path-Dependent Neural Jump ODEs
- A Posterior-Dynamics Framework for Imaging Inverse Problems with Pretrained Diffusion Priors
- Scalable Posterior Uncertainty for Flexible Density-Based Clustering
- A Transport Geometry of Benign Overfitting
- Efficient quadratic entropy with distance sketches