---
library_name: transformers
model_name: Phi-4-Argunaut-1-HIRPO
tags:
- logic
- argumentation
- critical-thinking
- argument-mapping
- generated_from_trainer
- trl
- rlvr
- hirpo
licence: license
datasets:
- DebateLabKIT/arguments-and-debates
language:
- en
base_model:
- DebateLabKIT/Phi-4-Argunaut-1-SPIN-dev1
pipeline_tag: text-generation
---

# Model Card for Phi-4-Argunaut-1-HIRPO

This model is a fine-tuned version of [DebateLabKIT/Phi-4-Argunaut-1-SPIN-dev1]([DebateLabKIT/Phi-4-Argunaut-1-SPIN-dev1](https://huggingface.co/DebateLabKIT/Phi-4-Argunaut-1-SPIN-dev1)).
It has been trained using [TRL](https://github.com/huggingface/trl).


📘 [HF Blog Article](https://huggingface.co/blog/ggbetz/argunauts-update-202512)

[<img src="https://raw.githubusercontent.com/wandb/assets/main/wandb-github-badge-28.svg" alt="Visualize in Weights & Biases" width="150" height="24"/>](https://api.wandb.ai/links/ggbetz/tqp681ch)



## Quick start

```python
from transformers import pipeline

question = "If you had a time machine, but could only go to the past or the future once and never return, which would you choose and why?"
generator = pipeline("text-generation", model="DebateLabKIT/Llama-3.1-Argunaut-1-8B-HIRPO", device="cuda")
output = generator([{"role": "user", "content": question}], max_new_tokens=128, return_full_text=False)[0]
print(output["generated_text"])
```

## Training procedure


This model was trained with Hindsight Instruction Relabeling Preference Optimization (HIRPO), an Online DPO version derived from [The Wisdom of Hindsight Makes Language Models Better Instruction Followers](https://huggingface.co/papers/2302.05206).

More details about the training procedure can be found in the [blog post](https://huggingface.co/blog/ggbetz/argunauts-phase-3).

We have released the preference pairs generated online as a separate dataset: [DebateLabKIT/argunauts-hirpo-preferences](https://huggingface.co/datasets/DebateLabKIT/argunauts-hirpo-preferences).

### Framework versions

- TRL: 0.19.1
- Transformers: 4.53.3
- Pytorch: 2.4.1
- Datasets: 3.1.0
- Tokenizers: 0.21.4

## Evaluation

As described in [this article](https://huggingface.co/blog/ggbetz/argunauts-update-202512), `Phi-4-Argunaut-1-HIRPO` technically masters formal argument analysis but has lost general conversational abilities during one-sided training.


## Citations

Cite HIR as:

```bibtex
@misc{zhang2023wisdomhindsightmakeslanguage,
      title={The Wisdom of Hindsight Makes Language Models Better Instruction Followers}, 
      author={Tianjun Zhang and Fangchen Liu and Justin Wong and Pieter Abbeel and Joseph E. Gonzalez},
      year={2023},
      eprint={2302.05206},
      archivePrefix={arXiv},
      primaryClass={cs.CL},
      url={https://arxiv.org/abs/2302.05206}, 
}
```

Cite TRL as:
    
```bibtex
@misc{vonwerra2022trl,
	title        = {{TRL: Transformer Reinforcement Learning}},
	author       = {Leandro von Werra and Younes Belkada and Lewis Tunstall and Edward Beeching and Tristan Thrush and Nathan Lambert and Shengyi Huang and Kashif Rasul and Quentin Gallouédec},
	year         = 2020,
	journal      = {GitHub repository},
	publisher    = {GitHub},
	howpublished = {\url{https://github.com/huggingface/trl}}
}
```