English
Related papers

Related papers: Recontextualization Mitigates Specification Gaming…

200 papers

Machine learning models are often brittle on production data despite achieving high accuracy on benchmark datasets. Benchmark datasets have traditionally served dual purposes: first, benchmarks offer a standard on which machine learning…

Machine Learning · Computer Science 2022-09-26 Matthew Groh

We propose a method for training language models in an interactive setting inspired by child language acquisition. In our setting, a speaker attempts to communicate some information to a listener in a single-turn dialogue and receives a…

Computation and Language · Computer Science 2025-05-12 Lennart Stöpler , Rufat Asadli , Mitja Nikolaus , Ryan Cotterell , Alex Warstadt

Language models (LMs) are susceptible to in-context reward hacking, where they exploit flaws in tainted or faulty written specifications or rubrics to achieve high scores without fulfilling the user's true intent. We introduce Specification…

Computation and Language · Computer Science 2025-07-28 Víctor Gallego

Transformer-based language models excel at in-context learning (ICL), where they can adapt to new tasks based on contextual examples, without parameter updates. In a specific form of ICL, which we refer to as \textit{contextual recall},…

Machine Learning · Computer Science 2026-03-24 Bhavya Vasudeva , Puneesh Deora , Alberto Bietti , Vatsal Sharan , Christos Thrampoulidis

Despite important progress, conversational systems often generate dialogues that sound unnatural to humans. We conjecture that the reason lies in their different training and testing conditions: agents are trained in a controlled "lab"…

Computation and Language · Computer Science 2021-04-01 Alberto Testoni , Raffaella Bernardi

Transformer-based language models are able to generate fluent text and be efficiently adapted across various natural language generation tasks. However, language models that are pretrained on large unlabeled web text corpora have been shown…

Computation and Language · Computer Science 2022-07-28 Farshid Faal , Ketra Schmitt , Jia Yuan Yu

Reinforcement learning from human feedback (RLHF) has emerged as a powerful technique to make large language models (LLMs) more capable in complex settings. RLHF proceeds as collecting human preference data, training a reward model on said…

Machine Learning · Computer Science 2024-02-05 Nathan Lambert , Roberto Calandra

We present symbol tuning - finetuning language models on in-context input-label pairs where natural language labels (e.g., "positive/negative sentiment") are replaced with arbitrary symbols (e.g., "foo/bar"). Symbol tuning leverages the…

Computation and Language · Computer Science 2024-01-02 Jerry Wei , Le Hou , Andrew Lampinen , Xiangning Chen , Da Huang , Yi Tay , Xinyun Chen , Yifeng Lu , Denny Zhou , Tengyu Ma , Quoc V. Le

The standard way to teach models is by feeding them lots of data. However, this approach often teaches models incorrect ideas because they pick up on misleading signals in the data. To prevent such misconceptions, we must necessarily…

Machine Learning · Computer Science 2024-08-23 Yoonho Lee , Michelle S. Lam , Helena Vasconcelos , Michael S. Bernstein , Chelsea Finn

We consider problems of making sequences of decisions to accomplish tasks, interacting via the medium of language. These problems are often tackled with reinforcement learning approaches. We find that these models do not generalize well…

Computation and Language · Computer Science 2020-10-07 Xusen Yin , Ralph Weischedel , Jonathan May

We study a linear contextual optimization problem where a decision maker has access to historical data and contextual features to learn a cost prediction model aimed at minimizing decision error. We adopt the predict-then-optimize framework…

Optimization and Control · Mathematics 2025-04-09 Omar Bennouna , Jiawei Zhang , Saurabh Amin , Asuman Ozdaglar

There is extensive interest in metric learning methods for image retrieval. Many metric learning loss functions focus on learning a correct ranking of training samples, but strongly overfit semantically inconsistent labels and require a…

Machine Learning · Computer Science 2023-06-05 Christopher Liao , Theodoros Tsiligkaridis , Brian Kulis

Consequential decision-making incentivizes individuals to strategically adapt their behavior to the specifics of the decision rule. While a long line of work has viewed strategic adaptation as gaming and attempted to mitigate its effects,…

Machine Learning · Computer Science 2020-02-19 John Miller , Smitha Milli , Moritz Hardt

Language models significantly benefit from context tokens, such as prompts or scratchpads. They perform better when prompted with informative instructions, and they acquire new reasoning capabilities by generating a scratch-pad before…

Computation and Language · Computer Science 2022-10-03 Charlie Snell , Dan Klein , Ruiqi Zhong

Despite the promising results of current cross-lingual models for spoken language understanding systems, they still suffer from imperfect cross-lingual representation alignments between the source and target languages, which makes the…

Computation and Language · Computer Science 2020-10-01 Zihan Liu , Genta Indra Winata , Peng Xu , Zhaojiang Lin , Pascale Fung

Recent reinforcement learning (RL) techniques have yielded impressive reasoning improvements in language models, yet it remains unclear whether post-training truly extends a model's reasoning ability beyond what it acquires during…

Computation and Language · Computer Science 2025-12-09 Charlie Zhang , Graham Neubig , Xiang Yue

Modern tokenizers employ deterministic algorithms to map text into a single "canonical" token sequence, yet the same string can be encoded as many non-canonical tokenizations using the tokenizer vocabulary. In this work, we investigate the…

Computation and Language · Computer Science 2026-02-04 Brian Siyuan Zheng , Alisa Liu , Orevaoghene Ahia , Jonathan Hayase , Yejin Choi , Noah A. Smith

Text classifiers are vulnerable to adversarial examples -- correctly-classified examples that are deliberately transformed to be misclassified while satisfying acceptability constraints. The conventional approach to finding adversarial…

Computation and Language · Computer Science 2024-05-21 Tom Roth , Inigo Jauregi Unanue , Alsharif Abuadbba , Massimo Piccardi

Prior works have demonstrated that implicit representations trained only for reconstruction tasks typically generate encodings that are not useful for semantic tasks. In this work, we propose a method that contextualises the encodings of…

Computer Vision and Pattern Recognition · Computer Science 2023-05-23 Theo W. Costain , Kejie Li , Victor A. Prisacariu

Large language models are able to learn new tasks in context, where they are provided with instructions and a few annotated examples. However, the effectiveness of in-context learning is dependent on the provided context, and the…

Computation and Language · Computer Science 2023-12-25 Afra Amini , Massimiliano Ciaramita