Probabilistic Reasoning with LLMs for Privacy Risk Estimation

Probabilistic reasoning is a key aspect of both human and artificial intelligence that allows for handling uncertainty and ambiguity in decision-making. In this paper, we introduce a new numerical reasoning task under uncertainty for large language models, focusing on estimating the privacy risk of user-generated documents containing privacy-sensitive information. We propose BRANCH, a new LLM methodology that estimates the k-privacy value of a text-the size of the population matching the given information. BRANCH factorizes a joint probability distribution of personal information as random variables. The probability of each factor in a population is estimated separately then combined to compute the final k-value using a Bayesian network. Our experiments show that this method successfully estimates the k-value 73% of the time, a 13% increase compared to o3-mini with chain-of-thought reasoning. We also find that LLM uncertainty is a good indicator for accuracy, as high variance predictions are 37.47% less accurate on average.

Based on the individual estimated answers, k = ⌈ 100,000 × 5% × (40% + 8%) × 20% × 10% ⌉ = 48 Reddit Post Other Documents Chatbot Message Personal Disclosure Detection Model ① Self-Disclosure Identification Python Interpreter ⑤ Probability Recombination Equation Generation Model Given the Bayesian model, the privacy risk is: k = ⌈ A × B × (C.1 + C.2) × D × E ⌉ User Document: Does Townsville have the highest inflation in the entire country? Been here 20 years. I work in Tech, but $10 for eggs is ridiculous! I don't have to deal with landlords and increasing rent, thankfully.

My daycare also increased their rate. I only have 4 months of maternity leave, so I'm looking for affordable child care options in the area.