See the sheer joy of my collaborators at #ACL2023 when 🤩
DissentQA
won best Paper AC award
This is a happy outcome of the fruitful collaboration with a group of wonderfully friendly people
https://arxiv.org/abs/2211.05655
#nlproc #machinelearning #Qa #factuality
#ACL2023 #nlproc #machinelearning #qa #factuality
False Promise of Imitating Proprietary LLMs: new paper argues that open source models imitating ChatGPT are less successful than perceived by human assessors because of imitating LLM style more successfully than content. Implies a need for more pre-training data, & improved evaluation data & method.
https://arxiv.org/pdf/2305.15717.pdf
#LLM #OpenSource #factuality #evaluation #assessment #chatgpt #AI #imitation #training #data #MachineLearning #safety #toxicity @machinelearning
#toxicity #safety #machinelearning #data #training #imitation #ai #chatgpt #assessment #evaluation #factuality #opensource #llm