JEE Advanced 25K: 25,000 Physics and Math Problems with Reasoning Traces

25,000 deduplicated questions for the Joint Entrance Examination (JEE) Advanced level, covering Physics and Mathematics. The dataset includes dynamic reasoning traces, topic-aligned metadata, and calculated intermediate arithmetic, created by author gss1147 and last updated on 2026-05-15. It features multiple problem families and contains no placeholder answers.

Use Cases

• Train or evaluate large language models on multi-step scientific reasoning based on the provided chain-of-thought traces.
• Benchmark AI performance on high-difficulty STEM problems based on the JEE Advanced difficulty level.
• Analyze pedagogical patterns and problem families within competitive exam preparation materials.
• Develop educational tools that provide step-by-step solutions based on the structured reasoning format.

Strengths

• Contains 25,000 questions, providing a substantial corpus for training.
• Includes dynamic chain-of-thought reasoning traces and calculated intermediate arithmetic.
• Questions are deduplicated and aligned with specific topics and a defined difficulty level (JEE Advanced).

Limitations

• Row count and file size are unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred from the provided format description.
• Freshness should be verified as the last update date is in the future (2026-05-15).

Provenance

Source

huggingface, author gss1147

Freshness

Last updated 2026-05-15 09:32:06