mlsanityData to 7 Oct 2026

Reinforcement Learning for LLM-based Event Forecasting

arXiv 2606.15917 ↗First posted 14 Jun 2026cs.LGcited by 2

On the map: LLM reasoning → Reinforcement learning for reasoning

Abstract (arXiv)

We use Group Relative Policy Optimization (GRPO), a recently devised sample and memory efficient reinforcement learning method, to finetune pretrained LLMs in the range of 1.5B to 14B parameters equipped with the ability to get current information through the use of a Wikipedia revisions tool, or news summaries, to forecast real events beyond the knowledge cutoff of the LLM, as well as problems made to simulate different aspects of the dynamics of that training. We use the results of these experiments to comment on the scaling capability of LLMs for forecasting, as well as classify how judgmental forecasting fits into the verifiable/unverifiable domain taxonomy, considering the impact of the inherent aleatoric uncertainty when forecasting future events (e.g. the roll of a die). As a result of the GRPO training, we manage to bring a 1.5B parameter transformer (Qwen 2.5 1.5B) to forecasting performance superior to Claude Sonnet 3.5 over the same dataset as measured by cross entropy from the market agreed probabilities. We also discuss various dead ends on the path to this result.

Key facts. Cited by 2 papers in the map (data through 7 Oct 2026). Placed in Reinforcement learning for reasoning, one of 17 directions in LLM reasoning. Builds on 11 of its 22 references that are in the map. First posted on arXiv 14 Jun 2026.

Builds on

11 of 22 references.

Closest ancestors · 1

Same direction, newest first

  1. 2025Outcome-based Reinforcement Learning to Predict the Future23 May

Other references · 10

  1. 2017Attention Is All You Needcited by 59,932
  2. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  3. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 8,851
  4. 2023A Comprehensive Survey of Continual Learning: Theory, Method and Applicationcited by 860
  5. 2024Approaching Human-Level Forecasting with Language Modelscited by 102
  6. 2022Forecasting Future World Events with Neural Networkscited by 63
  7. 2024Dated Data: Tracing Knowledge Cutoffs in Large Language Modelscited by 48
  8. 2024Consistency Checks for Language Model Forecasterscited by 27
  9. 2023AutoCast++: Enhancing World Event Prediction with Zero-shot Ranking-based Context Retrievalcited by 18
  10. 2025Evaluating LLMs on Real-World Forecasting Against Expert Forecasterscited by 6

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026LLM-based Agents for Forecasting and Prediction: Methods, Training, Evaluation, and Applicationscited by 0
  2. 2026Forecast-Dojo: Replayable Environments for Benchmarking and Training LLM Forecasting Agentscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Advancing Event Forecasting through Massive Training of Large Language Models: Challenges, Solutions, and Broader Impacts
  2. 2025Predictive Scaling Laws for Efficient GRPO Training of Large Reasoning Models
  3. 2026How Off-Policy Can GRPO Be? Mu-GRPO for Efficient LLM Reinforcement Learning
  4. 2025RiskPO: Risk-based Policy Optimization via Verifiable Reward for LLM Post-Training
  5. 2026Understanding Evolution Strategies for LLM Reasoning: Broader Reasoning Coverage than GRPO
  6. 2025Can GRPO Help LLMs Transcend Their Pretraining Origin?
  7. 2025Training-Free Group Relative Policy Optimization
  8. 2025Act Only When It Pays: Efficient Reinforcement Learning for LLM Reasoning via Selective Rollouts