Mind the Generation Process: Fine-Grained Confidence Estimation During LLM Generation

Open in new window