AdvisorQA: Towards Helpful and Harmless Advice-seeking Question Answering with Collective Intelligence

arXiv:2404.11826v1 Announce Type: new
Abstract: As the integration of large language models into daily life is on the rise, there is a clear gap in benchmarks for advising on subjective and personal dilemmas. To address this, we introduce AdvisorQA, the first benchmark developed to assess LLMs' capability in offering advice for deeply personalized concerns, utilizing the LifeProTips subreddit forum. This forum features a dynamic interaction where users post advice-seeking questions, receiving an average of 8.9 advice per query, with 164.2 …

abstract advice arxiv benchmark benchmarks capability clear collective cs.cl daily dilemmas gap integration intelligence language language models large language large language models life llms question question answering type