# Evaluation of Embedding-Based and Generative Methods for LLM-Driven Document Classification: Opportunities and Challenges

arXiv: https://arxiv.org/abs/2604.04997 · HTML: https://mlsanity.com/paper/2604.04997/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 8 of its 10 references that are in the map.
- First posted on arXiv 5 Apr 2026.

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

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [Gemma 3 Technical Report](https://mlsanity.com/paper/2503.19786/) (2025)
- [GME: Improving Universal Multimodal Retrieval by Multimodal LLMs](https://mlsanity.com/paper/2412.16855/) (2024)
- [Semantic Instance Segmentation with a Discriminative Loss Function](https://mlsanity.com/paper/1708.02551/) (2017)
- [mmE5: Improving Multimodal Multilingual Embeddings via High-quality Synthetic Data](https://mlsanity.com/paper/2502.08468/) (2025)
- [Improve Multi-Modal Embedding Learning via Explicit Hard Negative Gradient Amplifying](https://mlsanity.com/paper/2506.02020/) (2025)

## Abstract (arXiv)

This work presents a comparative analysis of embedding-based and generative models for classifying geoscience technical documents. Using a multi-disciplinary benchmark dataset, we evaluated the trade-offs between model accuracy, stability, and computational cost. We find that generative Vision-Language Models (VLMs) like Qwen2.5-VL, enhanced with Chain-of-Thought (CoT) prompting, achieve superior zero-shot accuracy (82%) compared to state-of-the-art multimodal embedding models like QQMM (63%). We also demonstrate that while supervised fine-tuning (SFT) can improve VLM performance, it is sensitive to training data imbalance.