LEAP: Likelihood Elicitation and Aggregation for LLM-based Probabilistic Forecasting

arXiv:2609.01337 · cs.AI · Submitted 2026-09-01 · Read on arXiv

cs.AI

Submitted: 2026-09-01

Updated: 2026-09-01

Comments: Accepted to the 2026 Conference on Empirical Methods in Natural Language Processing (EMNLP 2026). 20 pages, 3 figures, and 15 tables. Code: https://github.com/layingfish/LEAP

Code: https://github.com/layingfish/LEAP

License: http://creativecommons.org/licenses/by/4.0/

The gist: LLM-based forecasting systems have improved on real-world tasks such as financial markets and sports outcomes, largely through stronger search and tool use.

Terminology

Abstract

LLM-based forecasting systems have improved on real-world tasks such as financial markets and sports outcomes, largely through stronger search and tool use. Many systems still ask an LLM to read all collected evidence together and produce the final forecast. We call this design Monolithic Prediction. It can obscure how individual evidence items affect the result and collapse uncertainty across competing outcomes. We propose LEAP (Likelihood Elicitation and Aggregation for Probabilistic forecasting), which reorganizes how collected evidence is used in the prediction stage. LEAP examines each evidence item separately and elicits likelihood parameters that describe its implications for the target. An explicit prior and a deterministic probabilistic model then combine these likelihoods into a posterior distribution. This procedure supports continuous, single-choice, and multi-choice forecasts while preserving reproducible evidence contributions. We build a benchmark covering forecasting, information-seeking, and browsing tasks, and evaluate LEAP on our own agent loop and several agent CLI frameworks. Given the same evidence, LEAP improves most prediction and calibration metrics across models and remains stronger under controlled comparisons of prior access, inference budget, and aggregation.

Sources

Related papers