# An Invariant Learning Characterization of Controlled Text Generation

arXiv: https://arxiv.org/abs/2306.00198 · HTML: https://mlsanity.com/paper/2306.00198/

## Key facts

- Cited by 9 papers in the map (data through 14 Oct 2026).
- Builds on 60 of its 72 references that are in the map.
- First posted on arXiv 31 May 2023.

## Builds on (most cited first, 30 of 60)

- [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://mlsanity.com/paper/1810.04805/) (2018)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer](https://mlsanity.com/paper/1910.10683/) (2019)
- [Chain-of-Thought Prompting Elicits Reasoning in Large Language Models](https://mlsanity.com/paper/2201.11903/) (2022)
- [PaLM: Scaling Language Modeling with Pathways](https://mlsanity.com/paper/2204.02311/) (2022)
- [The Curious Case of Neural Text Degeneration](https://mlsanity.com/paper/1904.09751/) (2019)
- [Fine-Tuning Language Models from Human Preferences](https://mlsanity.com/paper/1909.08593/) (2019)
- [Invariant Risk Minimization](https://mlsanity.com/paper/1907.02893/) (2019)
- [Don't Stop Pretraining: Adapt Language Models to Domains and Tasks](https://mlsanity.com/paper/2004.10964/) (2020)
- [RealToxicityPrompts: Evaluating Neural Toxic Degeneration in Language Models](https://mlsanity.com/paper/2009.11462/) (2020)
- [LaMDA: Language Models for Dialog Applications](https://mlsanity.com/paper/2201.08239/) (2022)
- [Deep CORAL: Correlation Alignment for Deep Domain Adaptation](https://mlsanity.com/paper/1607.01719/) (2016)
- [In Search of Lost Domain Generalization](https://mlsanity.com/paper/2007.01434/) (2020)
- [Show Your Work: Scratchpads for Intermediate Computation with Language Models](https://mlsanity.com/paper/2112.00114/) (2021)
- [Plug and Play Language Models: A Simple Approach to Controlled Text Generation](https://mlsanity.com/paper/1912.02164/) (2019)
- [CTRL: A Conditional Transformer Language Model for Controllable Generation](https://mlsanity.com/paper/1909.05858/) (2019)
- [Out-of-Distribution Generalization via Risk Extrapolation (REx)](https://mlsanity.com/paper/2003.00688/) (2020)
- [Gender Bias in Coreference Resolution: Evaluation and Debiasing Methods](https://mlsanity.com/paper/1804.06876/) (2018)
- [Causal inference using invariant prediction: identification and confidence intervals](https://mlsanity.com/paper/1501.01332/) (2015)
- [Wizard of Wikipedia: Knowledge-Powered Conversational agents](https://mlsanity.com/paper/1811.01241/) (2018)
- [Underspecification Presents Challenges for Credibility in Modern Machine Learning](https://mlsanity.com/paper/2011.03395/) (2020)
- [DExperts: Decoding-Time Controlled Text Generation with Experts and Anti-Experts](https://mlsanity.com/paper/2105.03023/) (2021)
- [Nuanced Metrics for Measuring Unintended Bias with Real Data for Text Classification](https://mlsanity.com/paper/1903.04561/) (2019)
- [On Measuring Social Biases in Sentence Encoders](https://mlsanity.com/paper/1903.10561/) (2019)
- [Return of Frustratingly Easy Domain Adaptation](https://mlsanity.com/paper/1511.05547/) (2015)
- [GeDi: Generative Discriminator Guided Sequence Generation](https://mlsanity.com/paper/2009.06367/) (2020)
- [SeqGAN: Sequence Generative Adversarial Nets with Policy Gradient](https://mlsanity.com/paper/1609.05473/) (2016)
- [FUDGE: Controlled Text Generation With Future Discriminators](https://mlsanity.com/paper/2104.05218/) (2021)
- [Self-Diagnosis and Self-Debiasing: A Proposal for Reducing Corpus-Based Bias in NLP](https://mlsanity.com/paper/2103.00453/) (2021)
- [Measuring Bias in Contextualized Word Representations](https://mlsanity.com/paper/1906.07337/) (2019)

## Built on by (most cited first, 9 of 9)

- [On Behalf of the Stakeholders: Trends in NLP Model Interpretability in the Era of LLMs](https://mlsanity.com/paper/2407.19200/) (2024)
- [Text Generation: A Systematic Literature Review of Tasks, Evaluation, and Challenges](https://mlsanity.com/paper/2405.15604/) (2024)
- [Evaluating Alignment of Behavioral Dispositions in LLMs](https://mlsanity.com/paper/2602.11328/) (2026)
- [Automatic Construction of a Korean Toxic Instruction Dataset for Ethical Tuning of Large Language Models](https://mlsanity.com/paper/2311.18215/) (2023)
- [Actor Identification in Discourse: A Challenge for LLMs?](https://mlsanity.com/paper/2402.00620/) (2024)
- [Multi-environment Topic Models](https://mlsanity.com/paper/2410.24126/) (2024)
- [Linguistically-Controlled Paraphrase Generation](https://mlsanity.com/paper/2410.24199/) (2024)
- [First Place Solution of 2023 Global Artificial Intelligence Technology Innovation Competition Track 1](https://mlsanity.com/paper/2407.01271/) (2024)
- [Towards Context-Invariant Safety Alignment for Large Language Models](https://mlsanity.com/paper/2605.20994/) (2026)

## Abstract (arXiv)

Controlled generation refers to the problem of creating text that contains stylistic or semantic attributes of interest. Many approaches reduce this problem to training a predictor of the desired attribute. For example, researchers hoping to deploy a large language model to produce non-toxic content may use a toxicity classifier to filter generated text. In practice, the generated text to classify, which is determined by user prompts, may come from a wide range of distributions. In this paper, we show that the performance of controlled generation may be poor if the distributions of text in response to user prompts differ from the distribution the predictor was trained on. To address this problem, we cast controlled generation under distribution shift as an invariant learning problem: the most effective predictor should be invariant across multiple text environments. We then discuss a natural solution that arises from this characterization and propose heuristics for selecting natural environments. We study this characterization and the proposed method empirically using both synthetic and real data. Experiments demonstrate both the challenge of distribution shift in controlled generation and the potential of invariance methods in this setting.