Model Description
R1-Qwen-7B-R_HORIZON_Mixed1234 is a reasoning model used in the paper:
“R-Horizon: How Far Can Your Large Reasoning Model Really Go in Breadth and Depth?”
This model is trained on a mixed set of naively composed reasoning queries, including composition lengths n = 1, 2, 3, and 4, as introduced in the R-Horizon study.
The model is based on R1-distill-Qwen-7B and is designed to capture reasoning patterns across different composition depths, providing a more generalized reasoning capability compared to single-depth training models.
