LequeuISIR/AU-extraction_Qwen2.5-7B-Instruct

from vllm import LLM, SamplingParams

llm = LLM(model="LequeuISIR/AU-extraction_Qwen2.5-7B-Instruct", 
          max_model_len=2048)

tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2.5-7B-Instruct")

sampling_params = SamplingParams(temperature=0.2, max_tokens=2000)

messages = [
    {"role": "user", "content": f"{PROMPT}{item['text'].strip()}"}
]
prompt_string = tokenizer.apply_chat_template(
    messages, 
    tokenize=False, 
    add_generation_prompt=True
)

outputs = llm.generate([prompt_string], sampling_params)

Je vais te donner un texte d'opinion en français. Ton travail est de segmenter ce textes en unités argumentatives. \
Nous définissions une unité argumentative comme un ou des segments du texte qui s'intérèssent à un sujet. Elle peut être composée de solution(s), \
d'argument(s) ou de simple affirmation(s). Une unitée argumentative n'est pas forcément contigüe: elle peut joindre des segments qui ne se suivent pas. \n \
Cette tâche est EXTRACTIVE. tu dois COPIER et seulement copier le texte de l'unité argumentative exactement comme elle est écrite, incluant les majuscules et la ponctuation. \
Si l'unité argumentative est composée de plusieurs segments non contigues, tu peux les concaténer en les séparant simplement d'un espace. Il y a au moins \
une unité argumentative dans le texte, mais pas de nombre maximum. Ressors les unité argumentative sous forme de liste comme montré dans l'exemple. \
Il n'est pas obligatoire que tous les segments du texte fasse partie d'une unité argumentative. \n\n \

Tu dois donner les unités argumentative sous la forme du liste: \n \
- unité argumentative 1\n \
- unité argumentative 2\n \
...\n\n \

Ne sort RIEN D'AUTRE que la liste d'unités argumentatives.
voici le texte:\n \
"""

@inproceedings{lequeu-etal-2026-gdn,
    title = "The {GDN}-{CC} Dataset: Automatic Corpus Clarification for {AI}-enhanced Democratic Citizen Consultations",
    author = {Lequeu, Pierre-Antoine  and
      Labat, L{\'e}o  and
      Cave, Laur{\`e}ne  and
      Lejeune, Ga{\"e}l  and
      Yvon, Fran{\c{c}}ois  and
      Piwowarski, Benjamin},
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2026.acl-long.1523/",
    doi = "10.18653/v1/2026.acl-long.1523",
    pages = "32976--33006",
    ISBN = "979-8-89176-390-6",
    abstract = "LLMs are ubiquitous in modern NLP, and while their applicability extends to texts produced for democratic activities such as online deliberations or large-scale citizen consultations, ethical questions have been raised for their usage as analysis tools. We continue this line of research with two main goals: (a) to develop resources that can help standardize citizen contributions in public forums at the \textbf{pragmatic level}, and make them easier to use in topic modeling and political analysis; (b) to study how well this standardization can reliably be performed by small, open-weights LLMs, \textit{i.e.} models that can be run locally and transparently with limited resources. Accordingly, we introduce \textbf{Corpus Clarification} as a preprocessing framework for large-scale consultation data that transforms noisy, multi-topic contributions into structured, self-contained argumentative units ready for downstream analysis. We present \textbf{GDN-CC}, a manually-curated dataset of 1,231 contributions to the French \textit{Grand D{\'e}bat National}, comprising 2,285 argumentative units annotated for argumentative structure and manually clarified. We then show that finetuned Small Language Models match or outperform LLMs on reproducing these annotations, and measure their usability for an opinion clustering task. We finally release \textbf{GDN-CC-large}, an automatically annotated corpus of 240k contributions, the largest annotated democratic consultation dataset to date."
}