Abstract
An important aspect of developing LLMs that interact with humans is to align models' behavior to their users. It is possible to prompt an LLM into behaving as a certain persona, especially a user group or ideological persona the model captured during its pertaining stage. But, how to best align an LLM with a specific user and not a demographic or ideological group remains an open question. Mining public opinion surveys (by Pew Research), we find that the opinions of a user and their demographics and ideologies are not mutual predictors. We use this insight to align LLMs by modeling both user opinions as well as user demographics and ideology, achieving up to 7 points accuracy gains in predicting public opinions from survey questions across a broad set of topics. In addition to the typical approach of prompting LLMs with demographics and ideology, we discover that utilizing the most relevant past opinions from individual users enables the model to predict user opinions more accurately.
Abstract (translated)
开发与人类交互的LLM的重要方面是使其行为与用户对齐。可以通过prompt an LLM使其表现出某种人格特质,尤其是当模型在相关阶段捕获的用户组或意识形态人格特质。但是,如何最好地将LLM与特定的用户而不是 demographic 或意识形态团体对齐仍然是一个开放问题。通过Pew Research的公共意见调查进行数据挖掘,我们发现用户及其 demographic 和意识形态并不是相互预测的因素。利用这一洞察力,我们可以通过建模用户意见及其 demographic 和意识形态,对齐LLM,从广泛的主题上预测公共意见,最多可以提高7点的准确率。除了利用 demographic 和意识形态 prompts 的常见方法外,我们还发现,利用个体用户的最相关过去意见可以使模型更准确地预测用户意见。
URL
https://arxiv.org/abs/2305.14929