# Long Live Fine-Tuning: Task-Specific Transformers Outperform Zero-Shot LLMs for Misinformation Response Classification on Reddit

arXiv: https://arxiv.org/abs/2606.04274 · HTML: https://mlsanity.com/paper/2606.04274/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 24 of its 41 references that are in the map.
- First posted on arXiv 2 Jun 2026.

## Builds on (most cited first, 24 of 24)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [BART: Denoising Sequence-to-Sequence Pre-training for Natural Language Generation, Translation, and Comprehension](https://mlsanity.com/paper/1910.13461/) (2019)
- [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://mlsanity.com/paper/1910.01108/) (2019)
- [A Broad-Coverage Challenge Corpus for Sentence Understanding through Inference](https://mlsanity.com/paper/1704.05426/) (2017)
- [FEVER: a large-scale dataset for Fact Extraction and VERification](https://mlsanity.com/paper/1803.05355/) (2018)
- [Exploiting Cloze Questions for Few Shot Text Classification and Natural Language Inference](https://mlsanity.com/paper/2001.07676/) (2020)
- [EDA: Easy Data Augmentation Techniques for Boosting Performance on Text Classification Tasks](https://mlsanity.com/paper/1901.11196/) (2019)
- [A Survey on Automated Fact-Checking](https://mlsanity.com/paper/2108.11896/) (2021)
- [Fake News Detection on Social Media: A Data Mining Perspective](https://mlsanity.com/paper/1708.01967/) (2017)
- [Benchmarking Zero-shot Text Classification: Datasets, Evaluation and Entailment Approach](https://mlsanity.com/paper/1909.00161/) (2019)
- [A Survey on Recent Approaches for Natural Language Processing in Low-Resource Scenarios](https://mlsanity.com/paper/2010.12309/) (2020)
- [Stance Detection with Bidirectional Conditional Encoding](https://mlsanity.com/paper/1606.05464/) (2016)
- [SemEval-2017 Task 8: RumourEval: Determining rumour veracity and support for rumours](https://mlsanity.com/paper/1704.05972/) (2017)
- [RumourEval 2019: Determining Rumour Veracity and Support for Rumours](https://mlsanity.com/paper/1809.06683/) (2018)
- [A Survey on Stance Detection for Mis- and Disinformation Identification](https://mlsanity.com/paper/2103.00242/) (2021)
- [Analysing How People Orient to and Spread Rumours in Social Media by Looking at Conversational Threads](https://mlsanity.com/paper/1511.07487/) (2015)
- [Stance and Sentiment in Tweets](https://mlsanity.com/paper/1605.01655/) (2016)
- [CASCADE: Contextual Sarcasm Detection in Online Discussion Forums](https://mlsanity.com/paper/1805.06413/) (2018)
- [Turing at SemEval-2017 Task 8: Sequential Approach to Rumour Stance Classification with Branch-LSTM](https://mlsanity.com/paper/1704.07221/) (2017)
- [Political Bias and Factualness in News Sharing across more than 100,000 Online Communities](https://mlsanity.com/paper/2102.08537/) (2021)
- [Stance Detection with BERT Embeddings for Credibility Analysis of Information on Social Media](https://mlsanity.com/paper/2105.10272/) (2021)

## Built on by (most cited first, 1 of 1)

- [Reading Anxiety or Reading the Label? Comparing Fine-Tuned and Frontier Models for Anxiety Detection on Social Media](https://mlsanity.com/paper/2609.20847/) (2026)

## Abstract (arXiv)

As large language models (LLMs) become default tools for online information verification, an implicit assumption follows them: that scale and general capability are sufficient for nuanced classification of misinformation discourse. We test this assumption directly on 900 Reddit comments spanning three PolitiFact-verified misinformation claims (environment, health, immigration), labelled as belief (propagates the claim), fact-check (corrects it), or other. We compare nine models across three paradigms -- BART-MNLI, three Llama variants, three commercial frontier LLMs (Claude Haiku 4.5, Gemini Flash Lite 2.5, Claude Sonnet 4.6), and fine-tuned DistilBERT and RoBERTa -- under universal and topic-specific label schemas. The assumption does not hold. Fine-tuned RoBERTa reaches 0.62 macro-F1 against a best zero-shot result of 0.50 (Claude Haiku 4.5), at a fraction of the per-query cost; the supervised advantage is concentrated on the belief class, the implicit, affective category every zero-shot model under-detects. Scaling does not help: Llama-3-8B matches Llama-3-70B, and Claude Sonnet 4.6 underperforms the smaller Haiku under generic labels, collapsing belief detection to 0.17 and refusing outright on a subset of comments flagged as sensitive. This is a safety-alignment artefact, not a capacity limit. Label schema and topic jointly shape zero-shot performance, with the same model varying by more than 0.13 macro-F1 across topics under matched labels. In a verification context, where missing belief is the costlier error, task-specific fine-tuning remains the more reliable choice despite the proliferation of large generative models.