Debiasing Large Language Models via Adaptive Causal Prompting with Sketch-of-Thought
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_ | 1866909988621910016 |
|---|---|
| author | Li, Bowen Xu, Ziqi Ren, Jing Luo, Renqiang Zhang, Xikun Zhang, Xiuzhen Ren, Yongli Xia, Feng |
| author_facet | Li, Bowen Xu, Ziqi Ren, Jing Luo, Renqiang Zhang, Xikun Zhang, Xiuzhen Ren, Yongli Xia, Feng |
| contents | Despite notable advancements in prompting methods for Large Language Models (LLMs), such as Chain-of-Thought (CoT), existing strategies still suffer from excessive token usage and limited generalisability across diverse reasoning tasks. To address these limitations, we propose an Adaptive Causal Prompting with Sketch-of-Thought (ACPS) framework, which leverages structural causal models to infer the causal effect of a query on its answer and adaptively select an appropriate intervention (i.e., standard front-door and conditional front-door adjustments). This design enables generalisable causal reasoning across heterogeneous tasks without task-specific retraining. By replacing verbose CoT with concise Sketch-of-Thought, ACPS enables efficient reasoning that significantly reduces token usage and inference cost. Extensive experiments on multiple reasoning benchmarks and LLMs demonstrate that ACPS consistently outperforms existing prompting baselines in terms of accuracy, robustness, and computational efficiency. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2601_08108 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Debiasing Large Language Models via Adaptive Causal Prompting with Sketch-of-Thought Li, Bowen Xu, Ziqi Ren, Jing Luo, Renqiang Zhang, Xikun Zhang, Xiuzhen Ren, Yongli Xia, Feng Computation and Language Artificial Intelligence Despite notable advancements in prompting methods for Large Language Models (LLMs), such as Chain-of-Thought (CoT), existing strategies still suffer from excessive token usage and limited generalisability across diverse reasoning tasks. To address these limitations, we propose an Adaptive Causal Prompting with Sketch-of-Thought (ACPS) framework, which leverages structural causal models to infer the causal effect of a query on its answer and adaptively select an appropriate intervention (i.e., standard front-door and conditional front-door adjustments). This design enables generalisable causal reasoning across heterogeneous tasks without task-specific retraining. By replacing verbose CoT with concise Sketch-of-Thought, ACPS enables efficient reasoning that significantly reduces token usage and inference cost. Extensive experiments on multiple reasoning benchmarks and LLMs demonstrate that ACPS consistently outperforms existing prompting baselines in terms of accuracy, robustness, and computational efficiency. |
| title | Debiasing Large Language Models via Adaptive Causal Prompting with Sketch-of-Thought |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2601.08108 |