Get Started
Home
Topics
Search
Library
Research questionHow can we detect internally incoherent language-model forecasts before relying on them for consequential decisions?Language models may assign probabilities to related events that cannot consistently arise from a single probability distribution. This can undermine trust in their forecasts even before the relevant outcomes are observed.
AI
Alignment & Safety
Evaluation & Benchmarks
Finance
Natural Language Processing
Statistical Machine Learning
Latest papersRecent research connected to this question, newest first.Dutch Books for Language ModelsThe evidence concerns forecasts for events generated from stock-returns data. Coherence is assessed without outcome labels using the largest guaranteed Dutch-book profit computed with linear programming; the reported analysis also examines richer logical relationships among events and the effect of irrelevant contextual details.research paper · Sep 2, 2026
Related questions
How can evaluators distinguish missing knowledge from miscalibrated outputs in language models?How can noisy LLM conditionals support coherent probabilistic inference over structured variables without order-dependent generation?Can language models infer others’ mental states as social interactions evolve under unreliable information?How can we test whether language models causally use scientific mechanisms instead of answer-correlated shortcuts?