Papers
arxiv:2509.23595

Timber: Training-free Instruct Model Refining with Base via Effective Rank

Published on Sep 28
Authors:
,
,
,
,

Abstract

Timber, a training-free method, enhances the exploration capability of Instruct models by partially reverting them towards the Base model, improving Pass@k performance.

AI-generated summary

Post-training, which elicits a pretrained Base model into the corresponding Instruct model, is widely considered to be superficial. In this work, we first reinforce this hypothesis by providing novel quantitative evidence from the weight level that the effective rank (eRank) remains negligibly changed. However, this superficiality also suffers a critical trade-off, improving the exploitation capabilities at the cost of limiting its exploration. To tackle this issue, we propose Timber, a simple yet effective training-free method that enhances the exploration capability of the Instruct model while preserving its exploitation. The key insight is to partially revert Instruct towards the paired Base model by subtle yet targeted refinement of the weight deltas. Extensive experiments on Llama and Qwen series demonstrate that Timber consistently improves vanilla Instruct models, particularly on Pass@k performance. Our findings offer new insights into the post-training stage at the weight level and practical strategies to refine the Instruct model without training.

Community

Sign up or log in to comment

Models citing this paper 1

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2509.23595 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2509.23595 in a Space README.md to link it from this page.

Collections including this paper 1