Loading...
Loading...
Available on 1 platform
Sign in to view source links and access this dataset
A highly curated mixture of coding agent traces designed specifically for Direct Preference Optimization (DPO) of Mixture-of-Experts (MoE) architecture models. The dataset employs Strategic Language Weighting to route the best coding agent data to specific language paradigms, such as Python and Bash. Author 'el4' uploaded it to Hugging Face, with a last recorded update on 2026-07-19.
License is unknown, which may restrict usage.