7 papers
FactCorrector: A Graph-Inspired Approach to Long-Form Factuality Correction of Large Language Models
Javier Carnerero-Cano, Massimiliano Pronesti, Radu Marinescu +6
Large language models (LLMs) are widely used in knowledge-intensive applications but often generate factually incorrect responses. A promising approach to rectify these flaws is co…
FactReasoner: A Probabilistic Approach to Long-Form Factuality Assessment for Large Language Models
Radu Marinescu, Debarun Bhattacharjya, Junkyu Lee +5
Large language models (LLMs) have achieved remarkable success in generative tasks, yet they often fall short in ensuring the factual accuracy of their outputs, thus limiting their…
SIMBA UQ: Similarity-Based Aggregation for Uncertainty Quantification in Large Language Models
Debarun Bhattacharjya, Balaji Ganesan, Junkyu Lee +4
When does a large language model (LLM) know what it does not know? Uncertainty quantification (UQ) provides measures of uncertainty, such as an estimate of the confidence in an LLM…
Interpreting LLM-as-a-Judge Policies via Verifiable Global Explanations
Jasmina Gajcin, Erik Miehling, Rahul Nair +3
Using LLMs to evaluate text, that is, LLM-as-a-judge, is increasingly being used at scale to augment or even replace human annotations. As such, it is imperative that we understand…
Multilinear and Linear Programs for Partially Identifiable Queries in Quasi-Markovian Structural Causal Models
João P. Arroyo, João G. Rodrigues, Daniel Lawand +6
We investigate partially identifiable queries in a class of causal models. We focus on acyclic Structural Causal Models that are quasi-Markovian (that is, each endogenous variable…
Optimistic Exploration for Risk-Averse Constrained Reinforcement Learning
James McCarthy, Radu Marinescu, Elizabeth Daly +1
Risk-averse Constrained Reinforcement Learning (RaCRL) aims to learn policies that minimise the likelihood of rare and catastrophic constraint violations caused by an environment's…