GLTR: Statistical Detection and Visualization of Generated Text

The rapid improvement of language models has raised the specter of abuse of text generation systems. This progress motivates the development of simple methods for detecting generated text that can be used by and explained to non-experts. We develop GLTR, a tool to support humans in detecting whether a text was generated by a model. GLTR applies a suite of baseline statistical methods that can detect generation artifacts across common sampling schemes. In a human-subjects study, we show that the annotation scheme provided by GLTR improves the human detection-rate of fake text from 54% to 72% without any prior training. GLTR is open-source and publicly deployed, and has already been widely used to detect generated outputs

[1]  William Yang Wang.  “Liar, Liar Pants on Fire”: A New Benchmark Dataset for Fake News Detection , 2017, ACL.

[2]  Yann Dauphin,et al.  Hierarchical Neural Story Generation , 2018, ACL.

[3]  Daria Beresneva,et al.  Computer-Generated Text Detection Using Machine Learning: A Systematic Review , 2016, NLDB.

[4]  Ilya Sutskever,et al.  Language Models are Unsupervised Multitask Learners , 2019 .

[5]  Jason Weston,et al.  The Goldilocks Principle: Reading Children's Books with Explicit Memory Representations , 2015, ICLR.

[6]  Victor O. K. Li,et al.  Trainable Greedy Decoding for Neural Machine Translation , 2017, EMNLP.

[7]  Ming-Wei Chang,et al.  BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding , 2019, NAACL.

[8]  Navdeep Jaitly,et al.  Towards Better Decoding and Language Model Integration in Sequence to Sequence Models , 2016, INTERSPEECH.

[9]  Ido Dagan,et al.  Selective Sampling In Natural Language Learning , 1995 .

[10]  Scott Weinstein,et al.  Centering: A Framework for Modeling the Local Coherence of Discourse , 1995, CL.

[11]  Suhang Wang,et al.  Fake News Detection on Social Media: A Data Mining Perspective , 2017, SKDD.

[12]  François Yvon,et al.  Detecting Fake Content with Relative Entropy Scoring , 2008, PAN.

[13]  Massimo Poesio,et al.  Identifying fake Amazon reviews as learning from crowds , 2014, EACL.

[14]  Sameer Badaskar,et al.  Identifying Real or Fake Articles: Towards better Language Modeling , 2008, IJCNLP.

[15]  Percy Liang,et al.  Unifying Human and Statistical Evaluation for Natural Language Generation , 2019, NAACL.

[16]  Denny Britz,et al.  Generating High-Quality and Informative Conversation Responses with Sequence-to-Sequence Models , 2017, EMNLP.

[17]  Luke S. Zettlemoyer,et al.  Deep Contextualized Word Representations , 2018, NAACL.

[18]  Richard Socher,et al.  Learned in Translation: Contextualized Word Vectors , 2017, NIPS.

[19]  Dirk Hovy,et al.  The Enemy in Your Own Camp: How Well Can We Detect Statistically-Generated Fake Reviews – An Adversarial Study , 2016, ACL.

[20]  Ming Zhou,et al.  Machine Translation Detection from Monolingual Web-Text , 2013, ACL.