#semantic-diversity2 articles

2 articles

AIBullisharXiv – CS AI · 5d ago6/104

🧠

Post-training Large Language Models for Diverse High-Quality Responses

Researchers have developed DQO (Diversity Quality Optimization), a new training method that uses determinantal point processes to improve large language models' response diversity while maintaining quality. The approach addresses a key limitation of current reinforcement learning methods that tend to narrow LLM outputs to canonical responses.

AINeutralarXiv – CS AI · Feb 276/105

🧠

Evaluating the Diversity and Quality of LLM Generated Content

Research reveals that preference-tuned AI models like those using RLHF produce higher-quality diverse outputs than base models, despite appearing less diverse overall. The study introduces 'effective semantic diversity' metrics that account for quality thresholds, showing smaller models are more parameter-efficient at generating unique content.