Dan Klein
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Dan Klein has a total of 69 co-authors. Here are the top co-authors by number of shared publications:
- Kevin Yang
UC Berkeley
12 shared publications - Eve Fleisig
UC Berkeley, Berkeley, USA
8 shared publications - Ruiqi Zhong
University of California Berkeley
5 shared publications - Nicholas Tomlin
Computer Science Division UC Berkeley
4 shared publications - Yuandong Tian
Meta Superintelligence Labs (FAIR)
4 shared publications - Charlie Snell
University of California Berkeley
3 shared publications
Latest Publications
- Open Access
- 1 January 2025
- Published by Association for Computational Linguistics (ACL)
- p. 47-62
Balancing Quality and Variation: Spam Filtering Distorts Data Label Distributions
Eve Fleisig, Matthias Orlikowski, Philipp Cimiano, Dan Klein. Proceedings of the The 4th Workshop on Perspectivist Approaches to NLP. 2025.
- 1 January 2025
- Published by Association for Computational Linguistics (ACL)
- p. 47-62
Open AccessBalancing Quality and Variation: Spam Filtering Distorts Data Label Distributions
Eve Fleisig, Matthias Orlikowski, Philipp Cimiano, Dan Klein. Proceedings of the The 4th Workshop on Perspectivist Approaches to NLP. 2025.
- Open Access
- 1 January 2025
- Published by Association for Computational Linguistics (ACL)
- p. 7685-7711
ThoughtSculpt: Reasoning with Intermediate Revision and Search
Yizhou Chi, Kevin Yang, Dan Klein. Findings of the Association for Computational Linguistics: NAACL 2025. 2025.
- 1 January 2025
- Published by Association for Computational Linguistics (ACL)
- p. 7685-7711
Open AccessThoughtSculpt: Reasoning with Intermediate Revision and Search
Yizhou Chi, Kevin Yang, Dan Klein. Findings of the Association for Computational Linguistics: NAACL 2025. 2025.
- Open Access
Ghostbuster: Detecting Text Ghostwritten by Large Language Models
We introduce Ghostbuster, a state-of-the-art system for detecting AI-generated text. Our method works by passing documents through a series of weaker language models, running a structured search over possible combinations of their features, and then training a classifier on the selected features to predict whether documents are AI-generated. Crucially, Ghostbuster does not require access to token probabilities from the target model, making it useful for detecting text generated by black-box models or unknown model versions. In conjunction with our model, we release three new datasets of human- and AI-generated text as detection benchmarks in the domains of student essays, creative writing, and news articles. We compare Ghostbuster to a variety of existing detectors, including DetectGPT and GPTZero, as well as a new RoBERTa baseline. Ghostbuster achieves 99.0 F1 when evaluated across domains, which is 5.9 F1 higher than the best preexisting model. It also outperforms all previous approaches in generalization across writing domains (+7.5 F1), prompting strategies (+2.1 F1), and language models (+4.4 F1). We also analyze the robustness of our system to a variety of perturbations and paraphrasing attacks and evaluate its performance on documents written by non-native English speakers.
Open AccessGhostbuster: Detecting Text Ghostwritten by Large Language Models
We introduce Ghostbuster, a state-of-the-art system for detecting AI-generated text. Our method works by passing documents through a series of weaker language models, running a structured search over possible combinations of their features, and then training a classifier on the selected features to predict whether documents are AI-generated. Crucially, Ghostbuster does not require access to token probabilities from the target model, making it useful for detecting text generated by black-box models or unknown model versions. In conjunction with our model, we release three new datasets of human- and AI-generated text as detection benchmarks in the domains of student essays, creative writing, and news articles. We compare Ghostbuster to a variety of existing detectors, including DetectGPT and GPTZero, as well as a new RoBERTa baseline. Ghostbuster achieves 99.0 F1 when evaluated across domains, which is 5.9 F1 higher than the best preexisting model. It also outperforms all previous approaches in generalization across writing domains (+7.5 F1), prompting strategies (+2.1 F1), and language models (+4.4 F1). We also analyze the robustness of our system to a variety of perturbations and paraphrasing attacks and evaluate its performance on documents written by non-native English speakers.
- Open Access
RLCD: Reinforcement Learning from Contrastive Distillation for Language Model Alignment
We propose Reinforcement Learning from Contrastive Distillation (RLCD), a method for aligning language models to follow principles expressed in natural language (e.g., to be more harmless) without using human feedback. RLCD creates preference pairs from two contrasting model outputs, one using a positive prompt designed to encourage following the given principles, and one using a negative prompt designed to encourage violating them. Using two different prompts causes model outputs to be more differentiated on average, resulting in cleaner preference labels in the absence of human annotations. We then use the preference pairs to train a preference model, which is in turn used to improve a base unaligned language model via reinforcement learning. Empirically, RLCD outperforms RLAIF (Bai et al., 2022b) and context distillation (Huang et al., 2022) baselines across three diverse alignment tasks--harmlessness, helpfulness, and story outline generation--and when using both 7B and 30B model scales for simulating preference data.
Open AccessRLCD: Reinforcement Learning from Contrastive Distillation for Language Model Alignment
We propose Reinforcement Learning from Contrastive Distillation (RLCD), a method for aligning language models to follow principles expressed in natural language (e.g., to be more harmless) without using human feedback. RLCD creates preference pairs from two contrasting model outputs, one using a positive prompt designed to encourage following the given principles, and one using a negative prompt designed to encourage violating them. Using two different prompts causes model outputs to be more differentiated on average, resulting in cleaner preference labels in the absence of human annotations. We then use the preference pairs to train a preference model, which is in turn used to improve a base unaligned language model via reinforcement learning. Empirically, RLCD outperforms RLAIF (Bai et al., 2022b) and context distillation (Huang et al., 2022) baselines across three diverse alignment tasks--harmlessness, helpfulness, and story outline generation--and when using both 7B and 30B model scales for simulating preference data.
- Top Cited PapersOpen Access
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 13541-13564
Linguistic Bias in ChatGPT: Language Models Reinforce Dialect Discrimination
Eve Fleisig, Genevieve Smith, Madeline Bossi, Ishita Rustagi, Xavier Yin, Dan Klein. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024.
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 13541-13564
Top Cited PapersOpen AccessLinguistic Bias in ChatGPT: Language Models Reinforce Dialect Discrimination
Eve Fleisig, Genevieve Smith, Madeline Bossi, Ishita Rustagi, Xavier Yin, Dan Klein. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024.
- Open Access
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 7468-7484
What Evidence Do Language Models Find Convincing?
Alexander Wan, Eric Wallace, Dan Klein. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2024.
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 7468-7484
Open AccessWhat Evidence Do Language Models Find Convincing?
Alexander Wan, Eric Wallace, Dan Klein. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2024.
- Open Access
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 15715-15724
American Sign Language Handshapes Reflect Pressures for Communicative Efficiency
Kayo Yin, Terry Regier, Dan Klein. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2024.
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 15715-15724
Open AccessAmerican Sign Language Handshapes Reflect Pressures for Communicative Efficiency
Kayo Yin, Terry Regier, Dan Klein. Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2024.
- Open Access
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 2279-2292
The Perspectivist Paradigm Shift: Assumptions and Challenges of Capturing Human Labels
Eve Fleisig, Su Lin Blodgett, Dan Klein, Zeerak Talat. Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2024.
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 2279-2292
Open AccessThe Perspectivist Paradigm Shift: Assumptions and Challenges of Capturing Human Labels
Eve Fleisig, Su Lin Blodgett, Dan Klein, Zeerak Talat. Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2024.
- Top Cited PapersOpen Access
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 1702-1717
Ghostbuster: Detecting Text Ghostwritten by Large Language Models
Vivek Verma, Eve Fleisig, Nicholas Tomlin, Dan Klein. Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2024.
- 1 January 2024
- Published by Association for Computational Linguistics (ACL)
- p. 1702-1717
Top Cited PapersOpen AccessGhostbuster: Detecting Text Ghostwritten by Large Language Models
Vivek Verma, Eve Fleisig, Nicholas Tomlin, Dan Klein. Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). 2024.
- Open Access
Centering the Margins: Outlier-Based Identification of Harmed Populations in Toxicity Detection
The impact of AI models on marginalized communities has traditionally been measured by identifying performance differences between specified demographic subgroups. Though this approach aims to center vulnerable groups, it risks obscuring patterns of harm faced by intersectional subgroups or shared across multiple groups. To address this, we draw on theories of marginalization from disability studies and related disciplines, which state that people farther from the norm face greater adversity, to consider the "margins" in the domain of toxicity detection. We operationalize the "margins" of a dataset by employing outlier detection to identify text about people with demographic attributes distant from the "norm". We find that model performance is consistently worse for demographic outliers, with mean squared error (MSE) between outliers and non-outliers up to 70.4% worse across toxicity types. It is also worse for text outliers, with a MSE up to 68.4% higher for outliers than non-outliers. We also find text and demographic outliers to be particularly susceptible to errors in the classification of severe toxicity and identity attacks. Compared to analysis of disparities using traditional demographic breakdowns, we find that our outlier analysis frequently surfaces greater harms faced by a larger, more intersectional group, which suggests that outlier analysis is particularly beneficial for identifying harms against those groups.
Open AccessCentering the Margins: Outlier-Based Identification of Harmed Populations in Toxicity Detection
The impact of AI models on marginalized communities has traditionally been measured by identifying performance differences between specified demographic subgroups. Though this approach aims to center vulnerable groups, it risks obscuring patterns of harm faced by intersectional subgroups or shared across multiple groups. To address this, we draw on theories of marginalization from disability studies and related disciplines, which state that people farther from the norm face greater adversity, to consider the "margins" in the domain of toxicity detection. We operationalize the "margins" of a dataset by employing outlier detection to identify text about people with demographic attributes distant from the "norm". We find that model performance is consistently worse for demographic outliers, with mean squared error (MSE) between outliers and non-outliers up to 70.4% worse across toxicity types. It is also worse for text outliers, with a MSE up to 68.4% higher for outliers than non-outliers. We also find text and demographic outliers to be particularly susceptible to errors in the classification of severe toxicity and identity attacks. Compared to analysis of disparities using traditional demographic breakdowns, we find that our outlier analysis frequently surfaces greater harms faced by a larger, more intersectional group, which suggests that outlier analysis is particularly beneficial for identifying harms against those groups.