On Measuring Social Biases in Prompt-Based Learning
Large language models trained on a mixture of NLP tasks that are converted into a text-to-text format using prompts, can generalize into novel forms of language and handle novel tasks. A large body of work within prompt engineering attempts to understand the effects of input forms and prompts in achieving superior performance. We consider an alternative measure and inquire whether the way in which an input is encoded affects \textit{social biases} promoted in outputs. In this paper, we study T0, a large-scale multi-task text-to-text language model trained using prompt-based learning. We consider two different forms of semantically equivalent inputs: \textit{question-answer} format and \textit{premise-hypothesis} format. We use an existing bias benchmark for the former BBQ and create the first bias benchmark in natural language inference BBNLI with hand-written hypotheses while also converting each benchmark into the other form. The results on two benchmarks suggest that given two different formulations of essentially the same input, T0 conspicuously acts more biased in question answering form, which is seen during training, compared to premise-hypothesis form which is unlike its training examples.
Code (0)
등록된 구현이 없습니다.
Tasks
FormLanguage ModellingNatural Language InferencePrompt EngineeringQuestion AnsweringSimilar Papers 제목 키워드 기반
Challenges in Measuring Bias via Open-Ended Language Generation
Researchers have devised numerous ways to quantify social biases vested in pretrained language models. As some language models are capable of generating coherent completions given a set of textual prompts, several prompt…
Language ModelingLanguage ModellingText GenerationMeasuring Implicit Bias in Explicitly Unbiased Large Language Models
Large language models (LLMs) can pass explicit social bias tests but still harbor implicit biases, similar to humans who endorse egalitarian beliefs yet exhibit subtle biases. Measuring such implicit biases can be a chal…
Decision MakingDiagnosticLanguage ModellingBOLD: Dataset and Metrics for Measuring Biases in Open-Ended Language Generation
Recent advances in deep learning techniques have enabled machines to generate cohesive open-ended text when prompted with a sequence of words as context. While these models now empower many downstream applications from c…
BenchmarkingText GenerationBias Mitigation or Cultural Commonsense? Evaluating LLMs with a Japanese Dataset
Large language models (LLMs) exhibit social biases, prompting the development of various debiasing methods. However, debiasing methods may degrade the capabilities of LLMs. Previous research has evaluated the impact of b…
Are Large Language Models Chameleons? An Attempt to Simulate Social Surveys
Can large language models (LLMs) simulate social surveys? To answer this question, we conducted millions of simulations in which LLMs were asked to answer subjective questions. A comparison of different LLM responses wit…
Survey