To Burst or Not to Burst: Generating and Quantifying Improbable Text
Abstract: While LLMs are extremely capable at text generation, their outputs are still distinguishable from human-authored text. We explore this separation across many metrics over text, many sampling techniques, many types of text data, and across two popular LLMs, LLaMA and Vicuna. Along the way, we introduce a new metric, recoverability, to highlight differences between human and machine text; and we propose a new sampling technique, burst sampling, designed to close this gap. We find that LLaMA and Vicuna have distinct distributions under many of the metrics, and that this influences our results: Recoverability separates real from fake text better than any other metric when using LLaMA. When using Vicuna, burst sampling produces text which is distributionally closer to real text compared to other sampling techniques.
- A learning algorithm for boltzmann machines. Cognitive Science, 9(1):147–169.
- Language models are few-shot learners.
- Quantifying memorization across neural language models.
- Vicuna: An open-source chatbot impressing gpt-4 with 90%* chatgpt quality.
- Palm: Scaling language modeling with pathways.
- On the use of arxiv as a dataset.
- Hierarchical neural story generation. ArXiv, abs/1805.04833.
- Wikimedia Foundation. Wikimedia downloads.
- The Pile: An 800gb dataset of diverse text for language modeling. arXiv preprint arXiv:2101.00027.
- GLTR: Statistical detection and visualization of generated text. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics: System Demonstrations, pages 111–116, Florence, Italy. Association for Computational Linguistics.
- How efficiency shapes human language. Trends in Cognitive Sciences, 23(5):389–407.
- Alex Graves. 2012. Sequence transduction with recurrent neural networks.
- The curious case of neural text degeneration. In International Conference on Learning Representations.
- Scaling laws for neural language models.
- Ctrl: A conditional transformer language model for controllable generation.
- A watermark for large language models.
- A diversity-promoting objective function for neural conversation models.
- On the probability-quality paradox in language generation.
- DetectGPT: Zero-shot machine-generated text detection using probability curvature. In Proceedings of the 40th International Conference on Machine Learning, volume 202 of Proceedings of Machine Learning Research, pages 24950–24962. PMLR.
- Abstractive text summarization using sequence-to-sequence RNNs and beyond. In Proceedings of the 20th SIGNLL Conference on Computational Natural Language Learning, pages 280–290, Berlin, Germany. Association for Computational Linguistics.
- fairseq: A fast, extensible toolkit for sequence modeling. In Proceedings of NAACL-HLT 2019: Demonstrations.
- Improving language understanding by generative pre-training.
- Compressive transformers for long-range sequence modelling. In International Conference on Learning Representations.
- SemEval-2017 task 4: Sentiment analysis in Twitter. In Proceedings of the 11th International Workshop on Semantic Evaluation (SemEval-2017), pages 502–518, Vancouver, Canada. Association for Computational Linguistics.
- Release strategies and the social impacts of language models.
- Release strategies and the social impacts of language models. CoRR, abs/1908.09203.
- Edward Tian and Alexander Cui. 2023. Gptzero: Towards detection of ai-generated text using zero-shot and supervised methods.
- Llama: Open and efficient foundation language models. arXiv preprint arXiv:2302.13971.
- Attention is all you need.
- Opt: Open pre-trained transformer language models.
- Towards a unified multi-dimensional evaluator for text generation. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 2023–2038, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics.
- Texygen: A benchmarking platform for text generation models. In The 41st International ACM SIGIR Conference on Research & Development in Information Retrieval, SIGIR ’18, page 1097–1100, New York, NY, USA. Association for Computing Machinery.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.