Keyu He, Tejas Srinivasan, Brihi Joshi, Xiang Ren, Jesse Thomason, Swabha Swayamdipta. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Jaspreet Ranjit, Hyundong Justin Cho, Claire J. Smerdon, Yoonsoo Nam, Myles Phung, Jonathan May, John R. Blosnich, Swabha Swayamdipta. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2026.

Prior work has mapped which workplace tasks are exposed to AI, but less is known about whether workers perceive these tasks as meaningful or as busywork. We examined: (1) which dimensions of meaningful work do workers associate with tasks exposed to AI; and (2) how do the traits of existing AI systems compare to the traits workers want. We surveyed workers and developers on a representative sample of 171 tasks and use language models (LMs) to scale ratings to 10,131 computer-assisted tasks across all U.S. occupations. Worryingly, we find that tasks that workers associate with a sense of agency or happiness may be disproportionately exposed to AI. We also document design gaps: developers report emphasizing politeness, strictness, and imagination in system design; by contrast, workers prefer systems that are straightforward, tolerant, and practical. To address these gaps, we call for AI whose design explicitly focuses on meaningful work and worker needs, proposing a five-part research agenda.

Language model inversion seeks to recover hidden prompts using only language model outputs. This capability has implications for security and accountability in language model deployments, such as leaking private information from an API-protected language model’s system message. We propose a new method – prompt inversion from logprob sequences (PILS) – that recovers hidden prompts by gleaning clues from the model’s next-token probabilities over the course of multiple generation steps. Our method is enabled by a key insight: The vector-valued outputs of a language model occupy a low-dimensional subspace. This enables us to losslessly compress the full next-token probability distribution over multiple generation steps using a linear map, allowing more output information to be used for inversion. Our approach yields massive gains over previous state-of-the-art methods for recovering hidden prompts, achieving 2–3.5 times higher exact recovery rates across test sets, in one case increasing the recovery rate from 17% to 60%. Our method also exhibits surprisingly good generalization behavior; for instance, an inverter trained on 16 generations steps gets 5–27% higher prompt recovery when we increase the number of steps to 32 at test time. Furthermore, we demonstrate strong performance of our method on the more challenging task of recovering hidden system messages. We also analyze the role of verbatim repetition in prompt recovery and propose a new method for cross-family model transfer for logit-based inverters. Our findings suggest that next-token probabilities are a considerably more vulnerable attack surface for inversion attacks than previously known.

Jaspreet Ranjit, Brihi Joshi, Rebecca Dorn, Laura Petry, Olga Koumoundouros, Jayne Bottarini, Peichen Liu, Eric Rice, Swabha Swayamdipta. Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing. 2024.

Video summarization remains a huge challenge in computer vision due to the size of the input videos to be summarized. We propose an efficient, language-only video summarizer that achieves competitive accuracy with high data efficiency. Using only textual captions obtained via a zero-shot approach, we train a language transformer model and forego image representations. This method allows us to perform filtration amongst the representative text vectors and condense the sequence. With our approach, we gain explainability with natural language that comes easily for human interpretation and textual summaries of the videos. An ablation study that focuses on modality and data compression shows that leveraging text modality only effectively reduces input data processing while retaining comparable results.

Despite their ubiquity in language generation, it remains unknown why truncation sampling heuristics like nucleus sampling are so effective. We provide a theoretical explanation for the effectiveness of the truncation sampling by proving that truncation methods that discard tokens below some probability threshold (the most common type of truncation) can guarantee that all sampled tokens have nonzero true probability. However, thresholds are a coarse heuristic, and necessarily discard some tokens with nonzero true probability as well. In pursuit of a more precise sampling strategy, we show that we can leverage a known source of model errors, the softmax bottleneck, to prove that certain tokens have nonzero true probability, without relying on a threshold. Based on our findings, we develop an experimental truncation strategy and the present pilot studies demonstrating the promise of this type of algorithm. Our evaluations show that our method outperforms its threshold-based counterparts under automatic and human evaluation metrics for low-entropy (i.e., close to greedy) open-ended text generation. Our theoretical findings and pilot experiments provide both insight into why truncation sampling works, and make progress toward more expressive sampling algorithms that better surface the generative capabilities of large language models.

The hope of allowing programmers to more freely express themselves has led to a proliferation of program synthesis techniques. These tools automatically derive implementations from high-level specifications of user intent. These specifications may take the form of logical formulas, demonstrations, or input-output examples. Synthesizers guarantee that when synthesis is successful, the implementation satisfies the specification. However, they provide no additional information regarding how the implementation works or the manner in which the specification is realized. As a result, they remain algorithmic black boxes which are prone to producing unidiomatic code with procedurally generated identifier names, like $x 1, x 2$, etc. As a result, complicated implementations produced by modern program synthesizers are becoming increasingly hard to understand. One solution to this comprehensibility problem is to produce meaningful identifier names for its variables, functions, etc. While large language models (LLMs) suggest a simple way to obtain human-readable names, our experiments reveal that LLMs frequently produce nonsensical or misleading names when applied to code emitted by program synthesizers. In this paper, we develop an approach to reliably augment the implementation with explanatory names: We recover finegrained input-output data from the synthesis algorithm to enhance the prompt supplied to the LLM and use a combination of a program verifier and a second language model to validate the proposed names before presenting them to the user. Together, these techniques improve the accuracy of the proposed names from $\mathbf{2 4 \%}$ to $\mathbf{7 9 \%}$. A two-phase user study indicates that users significantly prefer the names produced by our technique, and that the proposed names greatly help users in understanding synthesized implementations.

Chandra Bhagavatula, Jena D. Hwang, Doug Downey, Ronan Le Bras, Ximing Lu, Lianhui Qin, Keisuke Sakaguchi, Swabha Swayamdipta, Peter West, Yejin Choi. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2023.

Hanjie Chen, Faeze Brahman, Xiang Ren, Yangfeng Ji, Yejin Choi, Swabha Swayamdipta. Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2023.

Alisa Liu, Zhaofeng Wu, Julian Michael, Alane Suhr, Peter West, Alexander Koller, Swabha Swayamdipta, Noah Smith, Yejin Choi. Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing. 2023.

Generative artificial intelligence has made significant strides, producing text indistinguishable from human prose and remarkably photorealistic images. Automatically measuring how close the generated data distribution is to the target distribution is central to diagnosing existing models and developing better ones. We present MAUVE, a family of comparison measures between pairs of distributions such as those encountered in the generative modeling of text or images. These scores are statistical summaries of divergence frontiers capturing two types of errors in generative modeling. We explore three approaches to statistically estimate these scores: vector quantization, non-parametric estimation, and classifier-based estimation. We provide statistical bounds for the vector quantization approach. Empirically, we find that the proposed scores paired with a range of $f$-divergences and statistical estimation methods can quantify the gaps between the distributions of human-written text and those of modern neural language models by correlating with human judgments and identifying known properties of the generated texts. We demonstrate in the vision domain that MAUVE can identify known properties of generated images on par with or better than existing metrics. In conclusion, we present practical recommendations for using MAUVE effectively with language and image modalities.

Maarten Sap, Swabha Swayamdipta, Laura Vianna, Xuhui Zhou, Yejin Choi, Noah Smith. Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2022.

Sarah Wiegreffe, Jack Hessel, Swabha Swayamdipta, Mark Riedl, Yejin Choi. Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. 2022.

Alisa Liu, Maarten Sap, Ximing Lu, Swabha Swayamdipta, Chandra Bhagavatula, Noah A. Smith, Yejin Choi. Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers). 2021.

Warning: this paper contains content that may be offensive or upsetting.Biased associations have been a challenge in the development of classifiers for detecting toxic language, hindering both fairness and accuracy.As potential solutions, we investigate recently introduced debiasing methods for text classification datasets and models, as applied to toxic language detection.Our focus is on lexical (e.g., swear words, slurs, identity mentions) and dialectal markers (specifically African American English).Our comprehensive experiments establish that existing methods are limited in their ability to prevent biased behavior in current toxicity detectors.We then propose an automatic, dialect-aware data correction method, as a proof-of-concept study.Despite the use of synthetic labels, this method reduces dialectal associations with toxicity.Overall, our findings show that debiasing a model trained on biased toxic language data is not as effective as simply relabeling the data to remove existing biases.

Contrastive explanations clarify why an event occurred in contrast to another.They are inherently intuitive to humans to both produce and comprehend.We propose a method to produce contrastive explanations in the latent space, via a projection of the input representation, such that only the features that differentiate two potential decisions are captured.Our modification allows model behavior to consider only contrastive reasoning, and uncover which aspects of the input are useful for and against particular decisions.Additionally, for a given input feature, our contrastive explanations can answer for which label, and against which alternative label, is the feature useful.We produce contrastive explanations via both highlevel abstract concept attribution and low-level input token/span attribution for two NLP classification benchmarks.Our findings demonstrate the ability of label-contrastive explanations to provide fine-grained interpretability of model decisions.1

As major progress is made in open-ended text generation, measuring how close machine-generated text is to human language remains a critical open problem. We introduce Mauve, a comparison measure for open-ended text generation, which directly compares the learnt distribution from a text generation model to the distribution of human-written text using divergence frontiers. Mauve scales up to modern text generation models by computing information divergences in a quantized embedding space. Through an extensive empirical study on three open-ended generation tasks, we find that Mauve identifies known properties of generated text, scales naturally with model size, and correlates with human judgments, with fewer restrictions than existing distributional evaluation metrics.

Language models pretrained on text from a wide variety of sources form the foundation of today's NLP. In light of the success of these broad-coverage models, we investigate whether it is still helpful to tailor a pretrained model to the domain of a target task. We present a study across four domains (biomedical and computer science publications, news, and reviews) and eight classification tasks, showing that a second phase of pretraining in-domain (domain-adaptive pretraining) leads to performance gains, under both high- and low-resource settings. Moreover, adapting to the task's unlabeled data (task-adaptive pretraining) improves performance even after domain-adaptive pretraining. Finally, we show that adapting to a task corpus augmented using simple data selection strategies is an effective alternative, especially when resources for domain-adaptive pretraining might be unavailable. Overall, we consistently find that multi-phase adaptive pretraining offers large gains in task performance.

As NLP models become larger, executing a trained model requires significant computational resources incurring monetary and environmental costs.To better respect a given inference budget, we propose a modification to contextual representation fine-tuning which, during inference, allows for an early (and fast) "exit" from neural network calculations for simple instances, and late (and accurate) exit for hard instances.To achieve this, we add classifiers to different layers of BERT and use their calibrated confidence scores to make early exit decisions.We test our proposed modification on five different datasets in two tasks: three text classification datasets and two natural language inference benchmarks.Our method presents a favorable speed/accuracy tradeoff in almost all cases, producing models which are up to five times faster than the state of the art, while preserving their accuracy.Our method also requires almost no additional training resources (in either time or parameters) compared to the baseline BERT model.Finally, our method alleviates the need for costly retraining of multiple models at different levels of efficiency; we allow users to control the inference speed/accuracy tradeoff using a single trained model, by setting a single variable at inference time.We publicly release our code.1 * Research completed during an internship at AI2. 1 github.com/allenai/sledgehammerLayer 0 Layer i Layer k Layer n Input Layer l Layer j Is confident?Yes

Swabha Swayamdipta, Roy Schwartz, Nicholas Lourie, Yizhong Wang, Hannaneh Hajishirzi, Noah A. Smith, Yejin Choi. Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP). 2020.

Large neural models have demonstrated human-level performance on language and vision benchmarks, while their performance degrades considerably on adversarial or out-of-distribution samples. This raises the question of whether these models have learned to solve a dataset rather than the underlying task by overfitting to spurious dataset biases. We investigate one recently proposed approach, AFLITE, which adversarially filters such dataset biases, as a means to mitigate the prevalent overestimation of machine performance. We provide a theoretical understanding for AFLITE, by situating it in the generalized framework for optimum bias reduction. We present extensive supporting evidence that AFLITE is broadly applicable for reduction of measurable dataset biases, and that models trained on the filtered datasets yield better generalization to out-of-distribution tasks. Finally, filtering results in a large drop in model performance (e.g., from 92% to 62% for SNLI), while human performance still remains high. Our work thus shows that such filtered datasets can pose new research challenges for robust generalization by serving as upgraded benchmarks.

We introduce the syntactic scaffold, an approach to incorporating syntactic information into semantic tasks.Syntactic scaffolds avoid expensive syntactic processing at runtime, only making use of a treebank during training, through a multitask objective.We improve over strong baselines on PropBank semantics, frame semantics, and coreference resolution, achieving competitive performance on all three tasks.

Hao Peng, Sam Thomson, Swabha Swayamdipta, Noah A. Smith. Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long Papers). 2018.

We present a transition-based parser that jointly produces syntactic and semantic dependencies. It learns a representation of the entire algorithm state, using stack long short-term memories. Our greedy inference algorithm has linear time, including feature extraction. On the CoNLL 2008--9 English shared tasks, we obtain the best published parsing performance among models that jointly learn syntax and semantics.

We describe a new dependency parser for English tweets, TWEEBOPARSER. The parser builds on several contributions: new syntactic annotations for a corpus of tweets (TWEEBANK), with conventions informed by the domain; adaptations to a statistical parsing algorithm; and a new approach to exploiting out-of-domain Penn Treebank data. Our experiments show that the parser achieves over 80% unlabeled attachment accuracy on our new, high-quality test set and measure the benefit of our contributions. Our dataset and parser can be found at http://www.ark.cs.cmu.edu/TweetNLP.