# Causal Inference for Human-Language Model Collaboration

arXiv: https://arxiv.org/abs/2404.00207 · HTML: https://mlsanity.com/paper/2404.00207/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Builds on 27 of its 59 references that are in the map.
- First posted on arXiv 30 Mar 2024.

## Builds on (most cited first, 27 of 27)

- [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)
- [PaLM: Scaling Language Modeling with Pathways](https://mlsanity.com/paper/2204.02311/) (2022)
- [Efficient Estimation of Word Representations in Vector Space](https://mlsanity.com/paper/1301.3781/) (2013)
- [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://mlsanity.com/paper/1910.01108/) (2019)
- [Variational Inference: A Review for Statisticians](https://mlsanity.com/paper/1601.00670/) (2016)
- [DialoGPT: Large-Scale Generative Pre-training for Conversational Response Generation](https://mlsanity.com/paper/1911.00536/) (2019)
- [Conditional Variational Autoencoder with Adversarial Learning for End-to-End Text-to-Speech](https://mlsanity.com/paper/2106.06103/) (2021)
- [Empirical Evaluation of Rectified Activations in Convolutional Network](https://mlsanity.com/paper/1505.00853/) (2015)
- [Causal Effect Inference with Deep Latent-Variable Models](https://mlsanity.com/paper/1705.08821/) (2017)
- [CoAuthor: Designing a Human-AI Collaborative Writing Dataset for Exploring Language Model Capabilities](https://mlsanity.com/paper/2201.06796/) (2022)
- [Learning Discourse-level Diversity for Neural Dialog Models using Conditional Variational Autoencoders](https://mlsanity.com/paper/1703.10960/) (2017)
- [Co-Writing Screenplays and Theatre Scripts with Language Models: An Evaluation by Industry Professionals](https://mlsanity.com/paper/2209.14958/) (2022)
- [Art and the science of generative AI: A deeper dive](https://mlsanity.com/paper/2306.04141/) (2023)
- [Desiderata for Representation Learning: A Causal Perspective](https://mlsanity.com/paper/2109.03795/) (2021)
- [Help me write a poem: Instruction Tuning as a Vehicle for Collaborative Poetry Writing](https://mlsanity.com/paper/2210.13669/) (2022)
- [A Generative Model For Zero Shot Learning Using Conditional Variational Autoencoders](https://mlsanity.com/paper/1709.00663/) (2017)
- [How to Make Causal Inferences Using Texts](https://mlsanity.com/paper/1802.02163/) (2018)
- [Understanding Iterative Revision from Human-Written Text](https://mlsanity.com/paper/2203.03802/) (2022)
- [Read, Revise, Repeat: A System Demonstration for Human-in-the-loop Iterative Text Revision](https://mlsanity.com/paper/2204.03685/) (2022)
- [Counterfactual Fairness with Disentangled Causal Effect Variational Autoencoder](https://mlsanity.com/paper/2011.11878/) (2020)
- [Plan, Write, and Revise: an Interactive System for Open-Domain Story Generation](https://mlsanity.com/paper/1904.02357/) (2019)
- [Human-Machine Collaboration Approaches to Build a Dialogue Dataset for Hate Speech Countering](https://mlsanity.com/paper/2211.03433/) (2022)
- [Causal Effects of Linguistic Properties](https://mlsanity.com/paper/2010.12919/) (2020)
- [Estimating Causal Effects of Tone in Online Debates](https://mlsanity.com/paper/1906.04177/) (2019)
- [Flexible text generation for counterfactual fairness probing](https://mlsanity.com/paper/2206.13757/) (2022)
- [Causal Estimation with Functional Confounders](https://mlsanity.com/paper/2102.08533/) (2021)

## Built on by (most cited first, 1 of 1)

- [Policy Learning with a Natural Language Action Space: A Causal Approach](https://mlsanity.com/paper/2502.17538/) (2025)

## Abstract (arXiv)

In this paper, we examine the collaborative dynamics between humans and language models (LMs), where the interactions typically involve LMs proposing text segments and humans editing or responding to these proposals. Productive engagement with LMs in such scenarios necessitates that humans discern effective text-based interaction strategies, such as editing and response styles, from historical human-LM interactions. This objective is inherently causal, driven by the counterfactual `what-if' question: how would the outcome of collaboration change if humans employed a different text editing/refinement strategy? A key challenge in answering this causal inference question is formulating an appropriate causal estimand: the conventional average treatment effect (ATE) estimand is inapplicable to text-based treatments due to their high dimensionality. To address this concern, we introduce a new causal estimand -- Incremental Stylistic Effect (ISE) -- which characterizes the average impact of infinitesimally shifting a text towards a specific style, such as increasing formality. We establish the conditions for the non-parametric identification of ISE. Building on this, we develop CausalCollab, an algorithm designed to estimate the ISE of various interaction strategies in dynamic human-LM collaborations. Our empirical investigations across three distinct human-LM collaboration scenarios reveal that CausalCollab effectively reduces confounding and significantly improves counterfactual estimation over a set of competitive baselines.