REVIEW 6 cited by
Quantifying Uncertainty in Answers from any Language Model and Enhancing their Trustworthiness
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
We introduce BSDetector, a method for detecting bad and speculative answers from a pretrained Large Language Model by estimating a numeric confidence score for any output it generated. Our uncertainty quantification technique works for any LLM accessible only via a black-box API, whose training data remains unknown. By expending a bit of extra computation, users of any LLM API can now get the same response as they would ordinarily, as well as a confidence estimate that cautions when not to trust this response. Experiments on both closed and open-form Question-Answer benchmarks reveal that BSDetector more accurately identifies incorrect LLM responses than alternative uncertainty estimation procedures (for both GPT-3 and ChatGPT). By sampling multiple responses from the LLM and considering the one with the highest confidence score, we can additionally obtain more accurate responses from the same LLM, without any extra training steps. In applications involving automated evaluation with LLMs, accounting for our confidence scores leads to more reliable evaluation in both human-in-the-loop and fully-automated settings (across both GPT 3.5 and 4).
Forward citations
Cited by 6 Pith papers
-
HalluField: Detecting LLM Hallucinations via Field-Theoretic Modeling
HalluField flags LLM hallucinations using a hand-weighted temperature-perturbation of token-level 'free energy' (negative log-likelihood) and Shannon entropy.
-
Conformal Language Model Reasoning with Coherent Factuality
A conformal prediction filter over deducibility graphs keeps language model reasoning coherently factual at user-chosen coverage levels on MATH and FELM problems.
-
Entropy Sentinel: Probing Entropy Traces for LLM Monitoring
Top-k decoding-entropy profiles can estimate and rank held-out domain accuracy for most tested LLMs, with difficulty-diverse training data the main success factor.
-
The Consistency Hypothesis in Uncertainty Quantification for Large Language Models
LLM generations that resemble other generations for the same question are more likely to be correct, and this rule can be used to build competitive black-box confidence scores.
-
From Calibration to Collaboration: LLM Uncertainty Quantification Should Be More Human-Centered
LLM uncertainty quantification should be judged by whether it improves real human decisions, not by calibration scores on trivia benchmarks.
-
ChemAU: Harness the Reasoning of LLMs in Chemical Research with Adaptive Uncertainty Estimation
ChemAU adds a position penalty to token-level uncertainty estimates so that flagged reasoning steps are corrected by a fine-tuned chemistry model, reporting improved accuracy on GPQA, MMLU-Pro, and SuperGPQA chemistry...
Discussion (0). Continue with ORCID to comment.