Skip to content

Dataset · evaleval

Every Eval Ever unified AI model evaluation database

evaleval/EEE_datastore

1 to 10 thousand JSON records with scores of various models from leaderboards, research papers, and local runs.

What's inside

An open database of benchmark results and AI model evaluations mapped to a unified schema. Contains 1,000 to 10,000 records in JSON format, gathered from public leaderboards, research papers, and local measurements. Each record corresponds to a single model evaluation run and can be supplemented with files containing results for individual examples.

What you can do with it

  • Comparative analysis and aggregation of metrics for different LLMs from independent sources without re-parsing
  • Building analytical dashboards and studying correlations between benchmarks
Tasks
Other
Size
1K–10K rows
Downloads this month
520.6K
Likes
63

How to load it

from datasets import load_dataset

ds = load_dataset("evaleval/EEE_datastore")