Compliance versus Sensibility: On the Reasoning Controllability in Large Language Models
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913165150781440 |
|---|---|
| author | Tan, Xingwei Valentino, Marco Akhter, Mahmud Elahi Zhou, Yuxiang Liakata, Maria Aletras, Nikolaos |
| author_facet | Tan, Xingwei Valentino, Marco Akhter, Mahmud Elahi Zhou, Yuxiang Liakata, Maria Aletras, Nikolaos |
| contents | Large Language Models (LLMs) are known to acquire reasoning capabilities through shared inference patterns in pre-training data, which are further elicited via Chain-of-Thought (CoT) practices. However, whether fundamental reasoning patterns, such as induction, deduction, and abduction, can be decoupled from specific problem instances remains a critical challenge for model controllability, and for shedding light on reasoning controllability. In this paper, we present the first systematic investigation of this problem through the lens of reasoning conflicts: an explicit tension between parametric and contextual information induced by mandating logical schemata that deviate from those expected for a target task. Our evaluation reveals that LLMs consistently prioritize sensibility over compliance, favoring task-appropriate reasoning patterns despite conflicting instructions. We further demonstrate that reasoning conflicts are internally detectable, as confidence scores significantly drop during conflicting episodes. Probing experiments confirm that reasoning types are linearly encoded from middle-to-late layers, indicating the potential for activation-level controllability. Leveraging these insights, we steer models towards compliance, increasing instruction following by up to 29%. Overall, our findings establish that while LLM reasoning is anchored to concrete instances, active mechanistic interventions can effectively decouple logical schemata from data, offering a path toward improved controllability, faithfulness, and generalizability. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_27251 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Compliance versus Sensibility: On the Reasoning Controllability in Large Language Models Tan, Xingwei Valentino, Marco Akhter, Mahmud Elahi Zhou, Yuxiang Liakata, Maria Aletras, Nikolaos Computation and Language Artificial Intelligence Large Language Models (LLMs) are known to acquire reasoning capabilities through shared inference patterns in pre-training data, which are further elicited via Chain-of-Thought (CoT) practices. However, whether fundamental reasoning patterns, such as induction, deduction, and abduction, can be decoupled from specific problem instances remains a critical challenge for model controllability, and for shedding light on reasoning controllability. In this paper, we present the first systematic investigation of this problem through the lens of reasoning conflicts: an explicit tension between parametric and contextual information induced by mandating logical schemata that deviate from those expected for a target task. Our evaluation reveals that LLMs consistently prioritize sensibility over compliance, favoring task-appropriate reasoning patterns despite conflicting instructions. We further demonstrate that reasoning conflicts are internally detectable, as confidence scores significantly drop during conflicting episodes. Probing experiments confirm that reasoning types are linearly encoded from middle-to-late layers, indicating the potential for activation-level controllability. Leveraging these insights, we steer models towards compliance, increasing instruction following by up to 29%. Overall, our findings establish that while LLM reasoning is anchored to concrete instances, active mechanistic interventions can effectively decouple logical schemata from data, offering a path toward improved controllability, faithfulness, and generalizability. |
| title | Compliance versus Sensibility: On the Reasoning Controllability in Large Language Models |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2604.27251 |