Bounding the Capabilities of Large Language Models in Open Text Generation with Prompt Constraints
Authors:
- Albert LuALAlbert Lu
- Yanzhe ZhangYZYanzhe Zhang
- Xuezhi WangXWXuezhi Wang
- Diyi YangDYDiyi Yang
Preprint
Abstract
The limits of open-ended generative models are unclear, yet increasingly important. What causes them to succeed and what causes them to fail? In this paper, we take a prompt-centric approach to analyzing and bounding the abilities of open-ended generative models. We present a generic methodology of analysis with two challenging prompt constraint types: structural and stylistic. These constraint types are categorized into a set of well-defined constraints that are analyzable by a single prompt. We then systematically create a diverse set of simple, natural, and useful prompts to robustly analyze each individual constraint. Using the GPT-3 text-davinci-002 model as a case study, we generate outputs from our collection of prompts and analyze the model's generative failures. We also show the generalizability of our proposed method on other large models like BLOOM and OPT. Our results and our in-context mitigation strategies reveal open challenges for future research. We have publicly released our code at https://github.com/SALT-NLP/Bound-Cap-LLM.Keywords
All Related Versions
This publication has 15 references indexed in Scilit:
- TruthfulQA: Measuring How Models Mimic Human Falsehoods2022
- Is GPT-3 Text Indistinguishable from Human Text? Scarecrow: A Framework for Scrutinizing Machine Text2022
- GLM: General Language Model Pretraining with Autoregressive Blank Infilling2022
- A Recipe for Arbitrary Text Style Transfer with Large Language Models2022
- Prompt Programming for Large Language Models: Beyond the Few-Shot Paradigm2021
- BOLD2021
- STORIUM: A Dataset and Evaluation Platform for Machine-in-the-Loop Story Generation2020
- DIALOGPT : Large-Scale Generative Pre-training for Conversational Response Generation2020
- Can a Suit of Armor Conduct Electricity? A New Dataset for Open Book Question Answering2018
- Automating readers' advisory to make book recommendations for K-12 readers2014