instruct-models

Tag

Cards List
#instruct-models

Drift-Constrained Optimization: Only Direction Matters in Fine-Tuning Instruct Models

arXiv cs.AI · 4d ago Cached

This paper proposes drift-constrained optimization for fine-tuning instruct models, focusing on the direction of updates to improve target-task performance while preserving existing capabilities. Experiments show that layer-selective probing can enhance scientific reasoning and multilingual translation across multiple languages.

0 favorites 0 likes
#instruct-models

Automatic or Controlled? Repetition Priming Reveals Divergent Processing in Base LLMs, Instruct LLMs, and Humans

arXiv cs.CL · 2026-08-18 Cached

The paper applies repetition priming to show that base LLMs use automatic processing while instruct models exhibit controlled processing, with humans displaying a hybrid profile.

0 favorites 0 likes
← Back to home

Submit Feedback