@inproceedings{chen-2026-deepgpt,
    title = "deepgpt at {S}em{E}val-2026 Task 1: A {C}hinese Humor Generation System via Instruction-Masked {QL}o{RA} and Reverse Constraint Data Mixing",
    author = "陈, 城",
    editor = "Kochmar, Ekaterina  and
      Ghosh, Debanjan  and
      North, Kai  and
      Komachi, Mamoru",
    booktitle = "Proceedings of the 20th {I}nternational {W}orkshop on {S}emantic {E}valuation (2026)",
    month = jul,
    year = "2026",
    address = "San Diego, California, USA",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl-workshops/2026.semeval-1.152/",
    pages = "1116--1121",
    ISBN = "979-8-89176-414-9",
    abstract = {AbstractThis paper presents the system description of the deepgpt team for SemEval2026 Task 1 (MWAHAHA: ComputationalHumor Generation), Subtask A. To address the challenge of generating highquality Chinese humor under strict textconstraints (e.g., incorporating speciffedrare words or relating to news headlines),we propose a parameter-e{\"i}{\textlnot}€{\"i}{\textlnot}€cient generation system based on Qwen2.5-3B-Instruct.We reconstructed 8,000 multi-source Chinese jokes into a conversational instruction tuning format. Crucially, to mitigate the prevalent issues of formatting hallucinations and template collapse, we introduce a strict Instruction Masking strategy during 4-bit QLoRA ffne-tuning. Bycompletely isolating the loss calculationto the target humorous text, the modelis forced to treat constraints as conditional inputs rather than conversationaldistributions to mimic. Empirical resultsshow that this architectural interventioncompletely eradicates meaningless conversational ffllers. Our system signiffcantlyboosted the hard constraint adherence (CAcc) to 94.6{\%} and achieved a highly competitive Elo rating of 903 in the o{\"i}{\textlnot}€{\"i}{\textlnot}€cialPairwise Human Evaluation, validating theeffectiveness of speciffc masking ffne-tuningfor lightweight large language models instrictly constrained generation tasks.}
}