← Back to explorer

SWE-ZERO-12M-trajectories

Type
dataset
Venue
AlienKevin
Year
2026
Source
huggingface
Access
free
Language
en
Added
2026-07-17T20:02:24.231940+00:00
Verified
2026-07-17T20:02:24.231940+00:00

Summary

The largest agentic-coding trace dataset to date: **112 B tokens** of execution-free agentic trajectories covering **122 K pull requests**, **3 K repositories**, and **16 programming languages**.

Keywords

hf-dataset language-modeling · -code parquet text datasets dask polars mlcroissant swe-zero code agentic pre-training

Topics

Language Modeling, Code

Research notes

  • downloads=12336; likes=124