File size: 1,949 Bytes
68a367e
 
 
 
 
 
 
 
 
 
 
 
cf7f1e6
68a367e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
---
license: apache-2.0
library_name: transformers
pipeline_tag: text-generation
tags:
- reasoning
- agent
- chain-of-thought
---

# ACTS: Agentic Chain-of-Thought Steering Controller

This repository contains a controller agent checkpoint for **ACTS (Agentic Chain-of-Thought Steering)**, presented in the paper [Agentic Chain-of-Thought Steering for Efficient and Controllable LLM Reasoning](https://huggingface.co/papers/2606.03965).

ACTS is a framework where a lightweight controller agent adaptively steers a frozen reasoner (such as DeepSeek-R1) step-by-step under a thinking-token budget. By formulating reasoning steering as a Markov decision process, the controller chooses a reasoning strategy and a short steering phrase at each step to enable controllable accuracy–efficiency trade-offs.

## Resources
- **Paper:** [Agentic Chain-of-Thought Steering for Efficient and Controllable LLM Reasoning](https://huggingface.co/papers/2606.03965)
- **Repository:** [Andree-9/ACTS](https://github.com/Andree-9/ACTS)
- **SFT Data:** [yuuxia/controller-sft-data](https://huggingface.co/datasets/yuuxia/controller-sft-data)

## Quick Start Inference

To use this controller to steer a reasoner, follow the setup instructions in the [GitHub repository](https://github.com/Andree-9/ACTS) and run the following command:

```bash
conda activate slime
./scripts/run_acts_inference.sh \
    --controller yuuxia/acts-controller \
    --reasoner   deepseek-ai/DeepSeek-R1-Distill-Qwen-7B \
    --benchmark  aime2024 \
    --budget     10000
```

## Citation

```bibtex
@misc{xia2026acts,
      title={Agentic Chain-of-Thought Steering for Efficient and Controllable LLM Reasoning},
      author={Yu Xia and Zhouhang Xie and Xin Xu and Byungkyu Kang and Prarit Lamba and Xiang Gao and Julian McAuley},
      year={2026},
      eprint={2606.03965},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2606.03965},
}
```