Impact of Decoding Methods on Human Alignment of Conversational LLMs
This work addresses the problem of aligning conversational LLMs with human speech for chatbot systems, though it is incremental as it focuses on optimizing existing decoding methods.
The paper investigated how decoding methods like Beam Search and Nucleus Sampling affect the alignment of LLM-generated conversations with human conversational norms, finding that fewer beams and lower P values improve alignment, with performance varying by dataset type.
To be included into chatbot systems, Large language models (LLMs) must be aligned with human conversational conventions. However, being trained mainly on web-scraped data gives existing LLMs a voice closer to informational text than actual human speech. In this paper, we examine the effect of decoding methods on the alignment between LLM-generated and human conversations, including Beam Search, Top K Sampling, and Nucleus Sampling. We present new measures of alignment in substance, style, and psychometric orientation, and experiment with two conversation datasets. Our results provide subtle insights: better alignment is attributed to fewer beams in Beam Search and lower values of P in Nucleus Sampling. We also find that task-oriented and open-ended datasets perform differently in terms of alignment, indicating the significance of taking into account the context of the interaction.