# When LLMs Can't Help: Real-World Evaluation of LLMs in Nutrition

arXiv: https://arxiv.org/abs/2511.20652 · HTML: https://mlsanity.com/paper/2511.20652/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 8 of its 69 references that are in the map.
- First posted on arXiv 7 Oct 2025.

## Builds on (most cited first, 8 of 8)

- [Prefix-Tuning: Optimizing Continuous Prompts for Generation](https://mlsanity.com/paper/2101.00190/) (2021)
- [Finetuned Language Models Are Zero-Shot Learners](https://mlsanity.com/paper/2109.01652/) (2021)
- [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://mlsanity.com/paper/1909.05858/) (2019)
- [A Survey of Evaluation Metrics Used for NLG Systems](https://mlsanity.com/paper/2008.12009/) (2020)
- [Baby Llama: knowledge distillation from an ensemble of teachers trained on a small dataset with no performance penalty](https://mlsanity.com/paper/2308.02019/) (2023)
- [ChatDiet: Empowering Personalized Nutrition-Oriented Food Recommender Chatbots through an LLM-Augmented Framework](https://mlsanity.com/paper/2403.00781/) (2024)
- [Can a Chatbot Determine My Diet?: Addressing Challenges of Chatbot Application for Meal Recommendation](https://mlsanity.com/paper/1802.09100/) (2018)
- [Comparing informativeness of an NLG chatbot vs graphical app in diet-information domain](https://mlsanity.com/paper/2206.13435/) (2022)

## Built on by (most cited first, 2 of 2)

- [COACH meets QUORUM: A Framework and Pipeline for Aligning User, Expert and Developer Perspectives in LLM-generated Health Counselling](https://mlsanity.com/paper/2603.08392/) (2026)
- [Quantifying Hidden Salt for Precision Healthcare: Sodium Assessment via Joint-Factor Retrieval and Chain-of-Thought Inference](https://mlsanity.com/paper/2609.22171/) (2026)

## Abstract (arXiv)

The increasing trust in large language models (LLMs), especially in the form of chatbots, is often undermined by the lack of their extrinsic evaluation. This holds particularly true in nutrition, where randomised controlled trials (RCTs) are the gold standard, and experts demand them for evidence-based deployment. LLMs have shown promising results in this field, but these are limited to intrinsic setups. We address this gap by running the first RCT involving LLMs for nutrition. We augment a rule-based chatbot with two LLM-based features: (1) message rephrasing for conversational variety and engagement, and (2) nutritional counselling through a fine-tuned model. In our seven-week RCT (n=81), we compare chatbot variants with and without LLM integration. We measure effects on dietary outcome, emotional well-being, and engagement. Despite our LLM-based features performing well in intrinsic evaluation, we find that they did not yield consistent benefits in real-world deployment. These results highlight critical gaps between intrinsic evaluations and real-world impact, emphasising the need for interdisciplinary, human-centred approaches.footnoteWe provide all of our code and results at:  hrefhttps://github.com/saeshyra/diet-chatbot-trialhttps://github.com/saeshyra/diet-chatbot-trial