# Large Language Models Struggle to Describe the Haystack without Human Help: Human-in-the-loop Evaluation of Topic Models

arXiv: https://arxiv.org/abs/2502.14748 · HTML: https://mlsanity.com/paper/2502.14748/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 20 of its 64 references that are in the map.
- First posted on arXiv 20 Feb 2025.

## Builds on (most cited first, 20 of 20)

- [Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks](https://mlsanity.com/paper/1908.10084/) (2019)
- [BERTScore: Evaluating Text Generation with BERT](https://mlsanity.com/paper/1904.09675/) (2019)
- [Measuring Mathematical Problem Solving With the MATH Dataset](https://mlsanity.com/paper/2103.03874/) (2021)
- [Dense Passage Retrieval for Open-Domain Question Answering](https://mlsanity.com/paper/2004.04906/) (2020)
- [HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering](https://mlsanity.com/paper/1809.09600/) (2018)
- [BERTopic: Neural topic modeling with a class-based TF-IDF procedure](https://mlsanity.com/paper/2203.05794/) (2022)
- [Regularizing and Optimizing LSTM Language Models](https://mlsanity.com/paper/1708.02182/) (2017)
- [Autoencoding Variational Inference For Topic Models](https://mlsanity.com/paper/1703.01488/) (2017)
- [Pre-training is a Hot Topic: Contextualized Document Embeddings Improve Topic Coherence](https://mlsanity.com/paper/2004.03974/) (2020)
- [Concept Induction: Analyzing Unstructured Text with High-Level Concepts Using LLooM](https://mlsanity.com/paper/2404.12259/) (2024)
- [Is Automated Topic Model Evaluation Broken?: The Incoherence of Coherence](https://mlsanity.com/paper/2107.02173/) (2021)
- [Probabilistic Latent Semantic Analysis](https://mlsanity.com/paper/1301.6705/) (2013)
- [Pathwise Derivatives Beyond the Reparameterization Trick](https://mlsanity.com/paper/1806.01851/) (2018)
- [Towards Understanding In-Context Learning with Contrastive Demonstrations and Saliency Maps](https://mlsanity.com/paper/2307.05052/) (2023)
- [PEDANTS: Cheap but Effective and Interpretable Answer Equivalence](https://mlsanity.com/paper/2402.11161/) (2024)
- [SODAPOP: Open-Ended Discovery of Social Biases in Social Commonsense Reasoning Models](https://mlsanity.com/paper/2210.07269/) (2022)
- [Investigating Crowdsourcing Protocols for Evaluating the Factual Consistency of Summaries](https://mlsanity.com/paper/2109.09195/) (2021)
- [SciDoc2Diagrammer-MAF: Towards Generation of Scientific Diagrams from Documents guided by Multi-Aspect Feedback Refinement](https://mlsanity.com/paper/2409.19242/) (2024)
- [Improving the TENOR of Labeling: Re-evaluating Topic Models for Content Analysis](https://mlsanity.com/paper/2401.16348/) (2024)
- [Discovering Mental Health Research Topics with Topic Modeling](https://mlsanity.com/paper/2308.13569/) (2023)

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

- [Spatial Reasoning in Multimodal Large Language Models: A Survey of Tasks, Benchmarks and Methods](https://mlsanity.com/paper/2511.15722/) (2025)
- [Compositional Coordination for Multi-Robot Teams with Large Language Models](https://mlsanity.com/paper/2507.16068/) (2025)

## Abstract (arXiv)

A common use of NLP is to facilitate the understanding of large document collections, with a shift from using traditional topic models to Large Language Models. Yet the effectiveness of using LLM for large corpus understanding in real-world applications remains under-explored. This study measures the knowledge users acquire with unsupervised, supervised LLM-based exploratory approaches or traditional topic models on two datasets. While LLM-based methods generate more human-readable topics and show higher average win probabilities than traditional models for data exploration, they produce overly generic topics for domain-specific datasets that do not easily allow users to learn much about the documents. Adding human supervision to the LLM generation process improves data exploration by mitigating hallucination and over-genericity but requires greater human effort. In contrast, traditional. models like Latent Dirichlet Allocation (LDA) remain effective for exploration but are less user-friendly. We show that LLMs struggle to describe the haystack of large corpora without human help, particularly domain-specific data, and face scaling and hallucination limitations due to context length constraints.