New VA-DPO method enables controllable emotion generation in language models

Researchers have developed a new method called VA-DPO to enable language models to generate text with controllable emotions. Unlike previous methods that use discrete labels, VA-DPO specifies desired affect as a continuous point in the Valence-Arousal plane. This approach modifies Direct Preference Optimization by using a frozen Valence-Arousal regressor to score generations and build preference data. Experiments show VA-DPO significantly reduces the distance to target emotions and improves correlation without negatively impacting performance on benchmarks like MMLU, HellaSwag, and TruthfulQA. AI

IMPACT Enables more nuanced and controllable emotional expression in AI-generated text.

RANK_REASON The cluster contains a research paper detailing a novel method for emotion generation in language models. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• Direct Preference Optimization
• HellaSwag
• Language Models
• Llama 3.1 8B-Instruct
• Llama 3.2:3b
• Massive Multitask Language Understanding
• Qwen3_8B
• TruthfulQA
• VA-DPO

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →