The paper measures LLM confidence per atomic fact, weights correctness by relevance, and uses high-confidence facts from the same response to correct low-confidence facts.
GPT-3.5-turbo (Brown et al., 2020) This model is part of OpenAI’s well-known GPT series
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Fact-Level Confidence Calibration and Self-Correction
The paper measures LLM confidence per atomic fact, weights correctness by relevance, and uses high-confidence facts from the same response to correct low-confidence facts.