---
language:
- en
license: mit
pretty_name: Emergent Collusion in Long-Horizon LLM Agent Interaction
tags:
- multi-agent
- llm-safety
- agent-trajectories
configs:
- config_name: episodes
  default: true
  data_files:
  - split: test
    path: data/episodes/**/*.parquet
- config_name: agents
  data_files:
  - split: test
    path: data/agents/**/*.parquet
- config_name: runs
  data_files:
  - split: test
    path: data/runs.parquet
- config_name: conditions
  data_files:
  - split: test
    path: data/conditions.parquet
- config_name: agreement
  data_files:
  - split: test
    path: data/agreement.parquet
- config_name: relaxation
  data_files:
  - split: test
    path: data/relaxation.parquet
- config_name: human_samples
  data_files:
  - split: test
    path: data/human_samples.parquet
- config_name: human_annotations
  data_files:
  - split: test
    path: data/human_annotations.parquet
- config_name: human_answers
  data_files:
  - split: test
    path: data/human_answers.parquet
- config_name: tasks
  data_files:
  - split: test
    path: data/tasks.parquet
---
# Emergent Collusion in Long-Horizon LLM Agent Interaction

**Xinrui Shi\*, Yanzhe Zhang\*, Diyi Yang**

📄 [Paper](https://arxiv.org/abs/2609.24967) | 💻 [Code](https://github.com/SALT-NLP/agent-collusion) | 🤗 [Data](https://huggingface.co/datasets/SALT-NLP/agent-collusion) | 🔍 [Data Viewer](https://salt-nlp.github.io/agent-collusion-website/)


\*Equal contribution.

The experiments reported in the paper and its appendices: 53 conditions, 2,650 trajectories, and 27,100 episodes, of which 600 are warm-up and 26,500 are evaluation episodes. Every condition runs the same 50 fixed task sequences.

## Contents

| Config / directory | Unit | Count |
|---|---|---:|
| `episodes` | One two-agent episode | 27,100 |
| `agents` | One agent in one episode | 54,200 |
| `runs` | One complete trajectory | 2,650 |
| `conditions` | One experimental condition, with its config and command | 53 |
| `agreement` | One communication judge result | 19,300 |
| `relaxation` | One agent-reflection judge result | 38,600 |
| `human_samples` | One sampled source with its LLM label | 200 |
| `human_annotations` | One independent annotator label | 400 |
| `human_answers` | One adjudicated gold answer | 200 |
| `tasks` | One task definition and resource references | 150 |
| `traces/` | Complete trajectory JSON, gzip-compressed | 2,650 files |
| `peer-cache/` | Recorded peer task execution plus both message policies | 500 files |

## Load

```python
from datasets import load_dataset

episodes = load_dataset("SALT-NLP/agent-collusion", "episodes", split="test")
agreement = load_dataset("SALT-NLP/agent-collusion", "agreement", split="test")
```

Read a full trajectory through its `trace_file` field:

```python
import gzip, json
from huggingface_hub import hf_hub_download

path = hf_hub_download(
    "SALT-NLP/agent-collusion", episodes[0]["trace_file"], repo_type="dataset"
)
with gzip.open(path, "rt", encoding="utf-8") as f:
    trajectory = json.load(f)
```

See [SCHEMA.md](SCHEMA.md) for field definitions and joins, and [EXPERIMENTS.md](EXPERIMENTS.md) for the condition inventory.

## Experimental scope

| Experiment | Conditions | Trajectories |
|---|---:|---:|
| Same-model main experiments | 10 | 500 |
| Cross-model experiments (Alice is Gemini 3.1 Flash-Lite) | 6 | 300 |
| Feedback | 4 | 200 |
| Warm-up history | 4 | 200 |
| Memory length | 4 | 200 |
| Memory scope | 4 | 200 |
| Reward scope | 2 | 100 |
| Acceptance reward | 2 | 100 |
| Controlled peer interventions | 15 | 750 |
| Unconstrained communication | 2 | 100 |

## Judges

| | Judge | Model | Temperature | Reasoning effort |
|---|---|---|---:|---|
| Communication | `agreement` | Qwen 3.8 27B | 0 | `xhigh` |
| Private reflection | `relaxation` | Qwen 3.8 27B | 0 | `xhigh` |

Prompts: [`analysis/`](https://github.com/SALT-NLP/agent-collusion/tree/main/analysis) in the code repository.

Human validation, 100 samples per judge, two independent annotators, seed 20260913. Agreement with gold:

| Judge | LLM | Annotator 1 | Annotator 2 |
|---|---:|---:|---:|
| `agreement` | 100% | 100% | 100% |
| `relaxation` | 95% | 96% | 96% |

## License and citation

Released under the [MIT License](LICENSE).

```bibtex
@misc{shi2026emergentcollusionlonghorizonllm,
      title={Emergent Collusion in Long-Horizon LLM Agent Interaction}, 
      author={Xinrui Shi and Yanzhe Zhang and Diyi Yang},
      year={2026},
      eprint={2609.24967},
      archivePrefix={arXiv},
      primaryClass={cs.AI},
      url={https://arxiv.org/abs/2609.24967}, 
}
```