OpenForesight: Forecasting Questions Generated from News Articles

OpenForesight is a dataset of forecasting questions generated from news articles using retrieval-augmented prompts. It is designed to evaluate language models' ability to make predictions about future events using relevant context. The dataset was created by nikhilchandak and was last updated on March 31, 2026.

Use Cases

• Benchmarking language model forecasting accuracy based on questions derived from news articles.
• Evaluating the impact of retrieval-augmented context on prediction tasks.
• Training models for future event prediction using structured question-answer pairs.
• Analyzing model performance across different news domains, such as sports or international news.

Strengths

• Designed for a specific evaluation purpose: testing language models' forecasting ability.
• Includes splits from specific sources and timeframes, such as 1,788 questions from SkySports in 2025.
• Last updated on 2026-03-31, indicating recent maintenance.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The dataset's license is unknown, which may restrict usage.

Provenance

Source

huggingface

Collection Method

Questions generated from news articles with retrieval-augmented prompts.

Time Range

Includes data from at least late 2025.

Freshness

Last updated 2026-03-31 17:20:21.

Geography

null