A Linguistic Comparison between Human and ChatGPT-Generated Conversations | allainews.com

Feb. 5, 2024, 6:48 a.m. | Morgan Sandler Hyesun Choung Arun Ross Prabu David

cs.CL updates on arXiv.org arxiv.org

This study explores linguistic differences between human and LLM-generated dialogues, using 19.5K dialogues generated by ChatGPT-3.5 as a companion to the EmpathicDialogues dataset. The research employs Linguistic Inquiry and Word Count (LIWC) analysis, comparing ChatGPT-generated conversations with human conversations across 118 linguistic categories. Results show greater variability and authenticity in human dialogues, but ChatGPT excels in categories such as social processes, analytical style, cognition, attentional focus, and positive emotional tone, reinforcing recent findings of LLMs being "more human than human." …

analysis authenticity chatgpt chatgpt-3.5 companion comparison conversations count cs.ai cs.cl cs.cy dataset differences generated human llm research show study word

More from arxiv.org / cs.CL updates on arXiv.org

STaR: Distilling Speech Temporal Relation for Lightweight Speech Self-Supervised Learning Models 1 day, 16 hours ago | arxiv.org

abstract arxiv computational cost +14

Large Language Models can Learn Rules 1 day, 16 hours ago | arxiv.org

abstract arxiv cs.ai cs.cl +18

Benchmarking LLMs via Uncertainty Quantification 1 day, 16 hours ago | arxiv.org

abstract arxiv benchmarking bridge +21

CARE: Extracting Experimental Findings From Clinical Literature 1 day, 16 hours ago | arxiv.org

abstract annotation applications arxiv +16

Prompt Cache: Modular Attention Reuse for Low-Latency Inference 1 day, 16 hours ago | arxiv.org

abstract arxiv attention cache +20

SpeechAlign: a Framework for Speech Translation Alignment Evaluation 1 day, 16 hours ago | arxiv.org

abstract advance alignment arxiv +14

I3: Intent-Introspective Retrieval Conditioned on Instructions 1 day, 16 hours ago | arxiv.org

abstract arxiv challenge cs.cl +10

DoDo Learning: DOmain-DemOgraphic Transfer in Language Models for Detecting Abuse Targeted at Public Figures 1 day, 16 hours ago | arxiv.org

abstract abuse arxiv automated +19

Investigating the prompt leakage effect and black-box defenses for multi-turn LLM interactions 1 day, 16 hours ago | arxiv.org

abstract arxiv box cs.ai +24

Data Architect

@ University of Texas at Austin | Austin, TX

View on ai-jobs.net

Data ETL Engineer

@ University of Texas at Austin | Austin, TX

View on ai-jobs.net

Lead GNSS Data Scientist

@ Lurra Systems | Melbourne

View on ai-jobs.net

Senior Machine Learning Engineer (MLOps)

@ Promaton | Remote, Europe

View on ai-jobs.net

Intern Large Language Models Planning (f/m/x)

@ BMW Group | Munich, DE

View on ai-jobs.net

Data Engineer Analytics

@ Meta | Menlo Park, CA | Remote, US

View on ai-jobs.net