Rank-1 LoRAs Encode Interpretable Reasoning Signals

2025-11-10

AI Summary

The internal mechanisms underlying performance improvements in reasoning models remain poorly understood. Method: We perform lightweight adaptation of Qwen-2.5-32B-Instruct using rank-1 LoRA, coupled with sparse autoencoder-based analysis of model activations, to uncover interpretable reasoning features embedded in low-rank adapters. Contribution: We demonstrate that minimal parameter perturbations—specifically, single-rank updates—are sufficient to elicit fine-grained, semantically homogeneous reasoning capabilities; their activation patterns are as interpretable as those of individual MLP neurons. On mainstream reasoning benchmarks (e.g., GSM8K, MMLU, HumanEval), our method recovers 73–90% of full fine-tuning performance. This work provides the first empirical evidence that complex reasoning abilities can be efficiently triggered by low-dimensional parameter changes, establishing a new paradigm for mechanistic interpretability that is both computationally efficient and highly interpretable.

Technology Category

• Knowledge Representation and Reasoning: Computational Complexity of Reasoning
• Machine Learning: Mixture of Experts (MoE)
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Reasoning models leverage inference-time compute to significantly enhance the performance of language models on difficult logical tasks, and have become a dominating paradigm in frontier LLMs. Despite their wide adoption, the mechanisms underpinning the enhanced performance of these reasoning models are not well understood. In this work, we show that the majority of new capabilities in reasoning models can be elicited by small, single-rank changes to base model parameters, with many of these changes being interpretable. Specifically, we use a rank-1 LoRA to create a minimal parameter adapter for Qwen-2.5-32B-Instruct which recovers 73-90% of reasoning-benchmark performance compared to a full parameter finetune. We find that the activations of this LoRA are as interpretable as MLP neurons, and fire for reasoning-specific behaviors. Finally, we train a sparse autoencoder on the entire activation state of this LoRA and identify fine-grained and monosemantic features. Our findings highlight that reasoning performance can arise largely from minimal changes to base model parameters, and explore what these changes affect. More broadly, our work shows that parameter-efficient training methods can be used as a targeted lens for uncovering fundamental insights about language model behavior and dynamics.

Problem

• Understanding mechanisms behind reasoning models' enhanced performance
• Demonstrating minimal parameter changes enable reasoning capabilities
• Using efficient training methods to interpret model behavior dynamics

Innovation

• Rank-1 LoRA enables minimal parameter adaptation
• LoRA activations interpret MLP-like reasoning behaviors
• Sparse autoencoder identifies fine-grained monosemantic features