A Multitask, Multilingual, Multimodal Evaluation of ChatGPT on Reasoning, Hallucination, and Interactivity

Abstract

This paper proposes a framework for quantitatively evaluating interactive LLMs such as ChatGPT using publicly available data sets. We carry out an extensive technical evaluation of ChatGPT using 23 data sets covering 8 different common NLP application tasks. We evaluate the multitask, multilingual and multi-modal aspects of ChatGPT based on these data sets and a newly designed multimodal dataset. We find that ChatGPT outperforms LLMs with zero-shot learning on most tasks and even outperforms fine-tuned models on some tasks. We find that it is better at understanding non-Latin script languages than generating them. It is able to generate multimodal content from textual prompts, via an intermediate code generation step. Moreover, we find that ChatGPT is 63.41% accurate on average in 10 different reasoning categories under logical reasoning, non-textual reasoning, and commonsense reasoning, hence making it an unreliable reasoner. It is, for example, better at deductive than inductive reasoning. ChatGPT suffers from hallucination problems like other LLMs and it generates more extrinsic hallucinations from its parametric memory as it does not have access to an external knowledge base. Finally, the interactive feature of ChatGPT enables human collaboration with the underlying LLM to improve its performance, i.e, 8% ROUGE-1 on summarization and 2% ChrF++ on machine translation, in a multi-turn "prompt engineering" fashion. We also release codebase for evaluation set extraction.

45 pages, AACL 2023

Cited by in corpus (21)

• Gender bias and stereotypes in Large Language Models
• Let's have a chat! A Conversation with ChatGPT: Technology, Applications, and Limitations
• Towards Interpretable Mental Health Analysis with Large Language Models
• Rethinking the Evaluation for Conversational Recommendation in the Era of Large Language Models
• Retrieving Supporting Evidence for Generative Question Answering
• The Perils & Promises of Fact-checking with Large Language Models
• ChatGPT Informed Graph Neural Network for Stock Movement Prediction
• Synergi: A Mixed-Initiative System for Scholarly Synthesis and Sensemaking
• CoAnnotating: Uncertainty-Guided Work Allocation between Human and Large Language Models for Data Annotation
• Selenite: Scaffolding Online Sensemaking with Comprehensive Overviews Elicited from Large Language Models
• WikiChat: Stopping the Hallucination of Large Language Model Chatbots by Few-Shot Grounding on Wikipedia
• Evaluating Large Language Models in Process Mining: Capabilities, Benchmarks, and Evaluation Strategies
• Large Knowledge Model: Perspectives and Challenges
• Computational Argumentation-based Chatbots: a Survey
• UZH_CLyp at SemEval-2023 Task 9: Head-First Fine-Tuning and ChatGPT Data Generation for Cross-Lingual Learning in Tweet Intimacy Prediction
• Learning to Describe for Predicting Zero-shot Drug-Drug Interactions
• Unmasking the giant: A comprehensive evaluation of ChatGPT's proficiency in coding algorithms and data structures
• Improving the Capabilities of Large Language Model Based Marketing Analytics Copilots With Semantic Search And Fine-Tuning
• Large Language Models and Multimodal Retrieval for Visual Word Sense Disambiguation
• Assessing Logical Puzzle Solving in Large Language Models: Insights from a Minesweeper Case Study
• Hallucinations in LLMs: A Lifecycle-Based Survey of Causes, Detection, Mitigation, and Prevention