Beyond Instruction Optimization: Multi-Agent Error-Driven Class Description Refinement for LLM-Based Classification
Hamvir Dev, Shivam Ratnakant Mhaskar, Sasanka Vutla, Anup Pattnaik
Abstract
Large Language Models (LLMs) have demonstrated considerable efficacy in classification tasks, yet their performance depends on two critical prompt components: Task Instructions (HOW to classify) and Class Descriptions (WHAT defines each class). While prompt engineering research has extensively explored instruction optimization, class descriptions have received comparatively less attention, often being treated as fixed inputs or simple label names. This represents a critical gap for real-world classification tasks, particularly in contact center domains, where labels often suffer from ambiguous boundaries, overlapping definitions, and incomplete coverage of possible cases—substantially limiting accuracy regardless of instruction quality.We propose a multi-agent framework for iteratively refining class descriptions based on classification errors. By analyzing misclassified instances, language agents automatically generate improved descriptions that better capture class distinctions and resolve ambiguities. Empirical evaluation across contact center and public benchmark datasets demonstrates upto 20.71% accuracy improvements over static class descriptions, addressing an orthogonal dimension to existing instruction optimization techniques.- Anthology ID:
- 2026.acl-industry.137
- Volume:
- Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026)
- Month:
- July
- Year:
- 2026
- Address:
- San Diego, California, USA
- Editors:
- Yunyao Li, Georg Rehm, Mei Tu
- Venue:
- ACL
- SIG:
- Publisher:
- Association for Computational Linguistics
- Note:
- Pages:
- 2031–2049
- Language:
- URL:
- https://preview.aclanthology.org/check-for-anonymous-pdfs/2026.acl-industry.137/
- DOI:
- Cite (ACL):
- Hamvir Dev, Shivam Ratnakant Mhaskar, Sasanka Vutla, and Anup Pattnaik. 2026. Beyond Instruction Optimization: Multi-Agent Error-Driven Class Description Refinement for LLM-Based Classification. In Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (ACL 2026), pages 2031–2049, San Diego, California, USA. Association for Computational Linguistics.
- Cite (Informal):
- Beyond Instruction Optimization: Multi-Agent Error-Driven Class Description Refinement for LLM-Based Classification (Dev et al., ACL 2026)
- PDF:
- https://preview.aclanthology.org/check-for-anonymous-pdfs/2026.acl-industry.137.pdf