Causal inference

Overview

This page covers only what the wiki's sources say about causal inference. There are eleven sources: a systematic review of model-agnostic counterfactual explanations 2103.04244 §1, a survey of causal machine learning 2206.15475 §1, a survey of causal inference in NLP that treats text as a confounder, an outcome or a treatment 2109.00725 §1, a position paper on neurosymbolic AI that discusses causal hierarchy 2012.05876 §5, a review of pitfalls in model-agnostic interpretation that has one section on causal interpretation 2007.04131 §9, a study of GPT-3 on classic cognitive-psychology tasks, one of which asks for interventions after passive observation 2206.14576 §Results / Causal reasoning, a randomised trial of an AI tutoring tool that estimates effects from random assignment 2410.03017 §Methods, a 2024 preprint on causal discovery with several LLM agents 2407.15073 §3.1, a 2020 position paper that proposes causal variables as inductive biases for deep learning 2011.15091 §3.5, a 2023 paper on using LLMs as simulated subjects, which argues that blinded simulations confound the treatment with invented covariates 2312.15524 §1, §2, and a 2026 paper that blames a failed LLM-agent run on a step by intervening on that step 2606.08275 §1. None is a general textbook on causal inference.

The review's central question is whether a counterfactual explanation can give causability, which it defines as an understanding that rests on causal relations rather than on feature correlations 2103.04244 §1.2. The methods share one core. They look for a nearby input x' that receives a different (target) prediction, by minimising a distance d(x, x') between x and x' 2103.04244 §4.1 eq. 4. The review argues that such post-hoc explainers capture correlations between features rather than the mechanism that produced the prediction 2103.04244 §1.2. Its conclusion is that the surveyed model-agnostic methods are not grounded in formal causal theory, so they may not give causability 2103.04244 §9.1 2103.04244 §10.

The causal ML survey supplies the formal vocabulary that the other sources assume. Its core is a structural causal model (SCM), in which each variable is a function of its parents and independent noise. In that setting, interventions and counterfactuals become answerable quantities 2206.15475 §2.3. Its point is that standard learning assumes test data follow the training distribution, and that causal assumptions are what make models hold up when it does not 2206.15475 §1.

The neurosymbolic position paper takes a different route to causal questions. It uses Pearl's three levels (association, intervention, counterfactual) and argues that a symbolic description of a trained network allows interventions and counterfactual questions 2012.05876 §5. The paper runs no experiments, so this is an argument, not a measured result [arxiv:2012.05876 §6].

The pitfalls review makes a narrower point. A model-agnostic importance score or effect plot describes what the model uses, and a predictive model exploits associations, so a non-zero importance does not show that a feature is a cause, an effect or unrelated 2007.04131 §9. The review says a partial dependence plot can be read as an average causal effect only if the model predicts well and the remaining variables form a valid adjustment set 2007.04131 §9. In its linear example, all five inputs get a similar non-zero effect, although three of them do not cause the outcome [arxiv:2007.04131 §9].

Formal core

The causal ML survey defines the terms used throughout this page 2206.15475 §2. The survey gives three layers of questions. The associational layer asks what seeing a value tells us about another variable. The interventional layer asks what happens when we set a variable. The counterfactual layer asks what would have happened to the same case under a different value 2206.15475 §2.2.2 2206.15475 §2.2.3. The counterfactual layer sits at the top, because it subsumes the two below it 2206.15475 §2.2.3.

An SCM is a set of structural assignments with independent noise terms, and each SCM induces a directed acyclic graph 2206.15475 §2.3. An interventional query keeps the noise prior. A counterfactual query has three steps: abduction (replace the noise prior by its posterior given the observation), action (replace the chosen equation), and prediction (rerun the equations on the same noise) 2206.15475 §2.3.2. The survey stresses that using the prior for a counterfactual gives the wrong answer 2206.15475 §2.3 2206.15475 §2.3.2.

A causal quantity is identifiable if it can be rewritten as a purely statistical one. Without hidden confounders and with a known graph, the backdoor criterion gives an adjustment set 2206.15475 §2.6. The survey does not detail the other graphical tests.

The NLP survey states the same core for a binary treatment. It defines the counterfactual outcome of a unit, the average treatment effect (ATE) and the conditional average treatment effect (CATE) 2109.00725 eq. 1 2109.00725 eq. 2. Three assumptions make the effect identifiable: conditional ignorability given the observed covariates 2109.00725 eq. 4, positivity, so that both treatment levels occur at every covariate value 2109.00725 eq. 5, and consistency 2109.00725 eq. 6. Under them, the counterfactual mean can be written as an expectation of observable conditional means 2109.00725 eq. 7. The survey adds that a causal DAG, read with d-separation, decides which variables to condition on 2109.00725 §2.3.

Variants

Counterfactual explanations (model-agnostic)

The review groups its algorithms into seven families by mechanism 2103.04244 §5. The paper's text says six categories in one place and lists seven, so the list below follows the list.

Instance-centric search

Changes the input features directly, under a loss and a distance. These follow the Lewis and Wachter counterfactual definition 2103.04244 §4.1 §6.1. Algorithms named: WatcherCF, Prototype, FACE, Weighted, TRUCE and DiCE. Card: Counterfactuals and Causability in Explainable Artificial Intelligence.

Constraint-centric search

Frames the search as a constraint satisfaction problem. MACE encodes the model, the distance and the constraints as one satisfiability formula, and asks a solver for a counterfactual within a distance threshold ε [arxiv:2103.04244 §6.2 eq. 20]. GIC and Coherent counterfactuals are also in this family [arxiv:2103.04244 §6.2]. Card: Counterfactuals and Causability in Explainable Artificial Intelligence.

Genetic-centric search

Uses crossover and mutation to search for counterfactuals. CERTIFAI scores robustness as the expected distance to the closest counterfactual [arxiv:2103.04244 §6.3 eq. 22]. MOCE optimises four objectives (prediction gap, Gower distance, number of changed features, and likelihood) with NSGA-II [arxiv:2103.04244 §6.3 eq. 23]. LORE generates a neighbourhood with a genetic algorithm and fits a decision tree to it 2103.04244 §6.3. Card: Counterfactuals and Causability in Explainable Artificial Intelligence.

Regression-centric search

Fits a local linear model to perturbed features, similar to LIME. LIME-C, SED-C and CLEAR belong here 2103.04244 §6.4. CLEAR scores its regression by a counterfactual fidelity error 2103.04244 §6.4 eq. 24. SED-C removes the smallest set of words that flips a linear text classifier [arxiv:2103.04244 §6.4]. Card: Counterfactuals and Causability in Explainable Artificial Intelligence.

Game-theory-centric search

Uses Shapley values, similar to SHAP. SHAP-C ranks features by Kernel SHAP and then runs the SED-C loop. SHAP-CC splits a contrastive "why P not Q" question into two Shapley-value questions 2103.04244 §6.5. Card: Counterfactuals and Causability in Explainable Artificial Intelligence.

Case-based reasoning

Retrieves a stored explanation case (a query and its good counterfactual), maps its features to the new query, and revises until the class flips [arxiv:2103.04244 §6.6]. Card: Counterfactuals and Causability in Explainable Artificial Intelligence.

Probabilistic search

Uses random walks, Markov sampling, VAEs or graphical models. The family contains PRINCE, C-CHVAE, MC-BRP, ABELE, CRUDS and RECOURSE [arxiv:2103.04244 §6.7]. RECOURSE is the only method the review marks as causal. It defines recourse actions as do-calculus interventions on a structural causal model that may be only partly known 2103.04244 §6.7 tab. 2. Card: Counterfactuals and Causability in Explainable Artificial Intelligence.

Symbolic extraction for interventional queries

Extract symbolic rules from a trained network, then answer causal questions over those rules. This is not a counterfactual search. The paper treats it as one route to explainability, since a rule set can give a proof trace where raw activations cannot 2012.05876 §5. It proposes that rules carry confidence values, so a system can signal when it does not know 2012.05876 §5. It names rule extraction from very large networks as an open challenge 2012.05876 §6. Card: Neurosymbolic AI: The 3rd Wave.

Causal machine learning (survey view)

The causal ML survey groups its methods into five problem areas. The entries below are the ones that matter for implementation. Its method summaries point to the primary papers, and it trains no models of its own 2206.15475 §1 2206.15475 §3. Card: Causal Machine Learning: A Survey and Open Problems.

Invariant risk minimization

Learns a feature map and one classifier that is optimal in every training environment. Its bilevel form and a Lagrangian relaxation are given 2206.15475 §3.2. The survey says it often performs no better than empirical risk minimisation, and that it works best with linear causal relations and enough environment heterogeneity 2206.15475 §3.2.

Causal generative models

Learns assignments for a causal graph, so that interventional and counterfactual samples can be drawn. DeepSCM learns each structural equation and inverts them for abduction; the survey notes it has been applied to graphs of order five 2206.15475 §4.1.1. Diff-SCM uses a diffusion model for abduction and classifier guidance for an intervention on the label 2206.15475 §4.2.

Causal feature attribution and recourse

Explanation methods that use a causal model rather than a predictive one. CXPlain trains an explanation model on causal influence measures [arxiv:2206.15475 §5.1.1]. Causal Shapley values replace conditioning on observations with conditioning by intervention 2206.15475 §5.1.4. Causal algorithmic recourse seeks the minimal-cost set of interventions that flips a prediction, with an SCM predicting their effects 2206.15475 §5.2.2.

Counterfactual fairness

A predictor is counterfactually fair if, for an individual, its prediction has the same distribution in the factual world and in the world where only the protected attribute is changed, with exogenous variables held fixed 2206.15475 §6.1. Path-specific variants count only effects along chosen unfair paths 2206.15475 §6.1.1. See algorithmic fairness.

Causal reinforcement learning

The survey lists eight RL problem types where a causal view is claimed to help, among them bandits, off-policy evaluation and credit assignment 2206.15475 §7. It reports these as claims from the cited work, and gives no comparison of its own.

Causal inference for text (NLP survey)

This variant is a map of two uses of causal reasoning with text. The first is estimating an effect when text plays a causal role. The second is making a text model robust, fair or explainable. The survey is a review and reports no experiments, so its claims are argued, not measured 2109.00725 §1 2109.00725 §5. Card: Causal Inference in Natural Language Processing: Estimation, Prediction, Interpretation and Beyond.

Text can play three roles in an effect estimate 2109.00725 §3. As a confounder, text carries variables that affect both treatment and outcome. The assumption at risk is that every confounder is measured. The survey says this cannot be tested from data and must rest on domain knowledge [arxiv:2109.00725 §3.1]. As an outcome, the measured quantity is often the output of an NLP model. The survey recommends fitting that model on a sample separate from the one used for the effect estimate 2109.00725 §3.2. As a treatment, the treatment is a latent property of the text 2109.00725 §3.3. Text can nearly determine its own treatment, so the overlap that adjustment needs (positivity) can fail in high-dimensional text 2109.00725 §3.1. The survey also names consistency failures, such as a hidden version of a treatment that is not recorded, and interference between units 2109.00725 §2.2 2109.00725 §3.2.

For robust and fair prediction 2109.00725 §4, the survey discusses counterfactual data augmentation (CDA). Manual edits are accurate but expensive. Keyword swaps can miss labels and covariates, and automatic rewriting may add new spurious correlations 2109.00725 §4.1.1. Distributional invariance criteria need labels for the nuisance factor or several labelled environments. The survey says they have been applied only to classification, and that many environments may be needed 2109.00725 §4.1.2. It also says invariance criteria are rarely applied to text in fairness work 2109.00725 §4.2. For interpretation, the survey casts an explanation as a comparison between an input and a counterfactual input, including counterfactual activations inside the model. It warns that attention weights can mislead and suggests intervening on internal nodes instead 2109.00725 §4.3.

The open problems include partial causal models, noisy proxies and structured outputs 2109.00725 §4.4. Heterogeneous effects and benchmarks are also open. The survey notes that no real-world text causal benchmark exists, because counterfactuals are never observed 2109.00725 §3.4. The survey gives no estimator for the identity in eq. 7. Regression adjustment, with one outcome model per treatment level, is the standard reading, and it is not the paper's method 2109.00725 eq. 7.

Causal discovery with LLM debate (MAC)

Causal discovery asks which variables have a direct causal effect on which, given observational data and optionally metadata 2407.15073 §3.1. Statistical methods use the data alone, and their validity depends on assumptions such as linearity, Gaussian noise, hidden confounders and sample size. No single algorithm is best on every dataset 2407.15073 §1 2407.15073 §3.4.

MAC uses two debate stages. In the first, agents choose a statistical algorithm from a pool, and a coding agent runs it on the full data to give an initial graph 2407.15073 §3.4 2407.15073 alg. 2. In the second, the graph is turned into plain-language constraints, added to the metadata, and refined by another debate 2407.15073 §3.5 2407.15073 alg. 3. Each debate has an Affirmative, a Negative and a Judge agent 2407.15073 §3.3. The paper's stated motivation is that a single LLM can repeat memorised associations and gives no challenge to a wrong answer 2407.15073 §3.2.

The results are author-reported. With Gemini-2.0-Flash, MAC comes first on 10 of 15 evaluation points across five benchmarks 2407.15073 §4.2. On Earthquake, exact search gets the same perfect scores [arxiv:2407.15073 tab. 2]. The single-agent version is competitive on most datasets at about half the cost, while the full version is much better on the harder graphs 2407.15073 §4.4 tab. 6. The full system uses 16 to 25 times the tokens of the standard baselines 2407.15073 §4.4. The agent prompts appear only as figures, and the card records no official code, so an implementation must write its own prompts and output parser 2407.15073 app. A.1. Card: Multi-Agent Causal Discovery Using Large Language Models.

Causal variables as an inductive bias (position paper)

This is a design argument, not a method. The paper proposes that the variables a learner uses should be high-level and verbalisable, and that these variables are often causal 2011.15091 §3.4 2011.15091 §3.5. Changes in the world should be sparse and local, starting at one or a few nodes of the causal graph 2011.15091 §3.6. An intervention on one mechanism should leave the others unchanged 2011.15091 §4.1. The paper also argues that shuffling data into an i.i.d. sample discards causal information that lies in the changes between sources 2011.15091 §4.2.

The only causal-learning example is a meta-learning setting taken from cited work. That work suggests that parameterising the correct variables lets a model adapt faster to interventions. The paper gives no numbers for this 2011.15091 §4.3. It gives no training procedure, and it calls combining these biases in one system an open question 2011.15091 §1. Card: Inductive Biases for Deep Learning of Higher-Level Cognition.

Causal reasoning probe for language models (wine casks)

This variant tests whether a language model answers interventional questions from observed data. It does not estimate an effect. The model sees 20 casks, each with the presence or absence of substances A, B and C, plus a stated causal structure. It is then asked how many of 20 new casks will contain C, under an observation (B is known to be present) or an intervention (B is added by hand) 2206.14576 §Results / Causal reasoning 2206.14576 fig. 5A.

The normative answers for the two questions differ only in the common-cause condition, where intervention cuts the arrows into B [arxiv:2206.14576 §Results / Causal reasoning]. In the causal-chain condition the normative answers are the same 2206.14576 fig. 5C. The paper reports that GPT-3 (Davinci) matches the normative intervention answers in the common-cause condition only when B is present. When B is absent, its intervention answers are not reduced 2206.14576 fig. 5B. In the causal-chain condition its answers are identical to those of the common-cause condition 2206.14576 fig. 5C. The paper's headline is that GPT-3 fails the causal task [arxiv:2206.14576 §Discussion]. All results are author-reported, and the causal figure's values are typed into the plotting script rather than computed from the outputs analyze.py. Card: Using cognitive psychology to understand GPT-3.

Randomised assignment and instrumental variables (trial analysis)

This family estimates an effect from an experiment, not from a model's predictions. Tutor CoPilot randomised tutors, not students, to get the tool or not 2410.03017 §Methods. Its primary model is an intent-to-treat regression: a session outcome on a treatment indicator, student covariates and a school-by-grade fixed effect, with the residual clustered by student-tutor pair 2410.03017 §Methods eq. 1. Under this design the coefficient estimates the effect of access. The paper reports a 4.0 percentage-point rise in the unconditional exit-ticket pass rate (standard error 0.015, control mean 0.617, Romano-Wolf adjusted p = 0.010) 2410.03017 tab. 1.

For the effect of use, the paper uses assignment as an instrument and estimates it by two-stage least squares. It reports a 14 percentage-point increase, from 62% to 76% 2410.03017 §Results. The card does not restate the instrument's exclusion assumptions, so check the paper's supplementary analysis before relying on this number. The trial ran for about two months, and students matched to a tutor with access had little exposure. The paper lists this as a limitation, and it may explain the null survey and end-of-year results 2410.03017 §Limitations. All results are author-reported.

Disclosing the randomisation to a simulated subject (LLM simulation)

This variant changes the prompt of an LLM that plays a subject in a price experiment. The study estimates demand for 40 consumer packaged-goods products against a human survey, with the price as the treatment 2312.15524 §2. In the usual blinded setup, each simulated customer sees one scenario and is not told that it is part of a test 2312.15524 §5. When the price changes, the model also changes the details it invents, such as the last price paid and a competing price 2312.15524 §2, fig. 1. The simulated demand curve is then inverted-U, while the human curve slopes downward 2312.15524 §2, fig. 2.

Adding covariates to the prompt was the first remedy the paper tests. Demographic personas, and then a sequence of twelve stages of behavioural covariates, were added to the prompt 2312.15524 §3, app. E. The error on the survey was lowest at stage 2 (MAE 0.0971, against 0.2250 at stage 1) and rose for most later stages 2312.15524 app. E. The paper reports a cost of the remedy, which it calls focalism: the model gives the stated details too much weight 2312.15524 §1, §3.

The second remedy is unblinding. Its system prompt says that the price is drawn at random and uniformly from a range, and that the customer sees one price and does not know about the draw 2312.15524 §5. The range is not given in the text; the minimum and maximum are placeholders 2312.15524 §5. Out of the box, GPT-4o-mini's mean absolute error (MAE) against the survey falls from 0.532 with the blinded prompt to 0.397 with the unblinded one 2312.15524 tab. 1. The text says unblinding lowered MAE for every model tested, with relative gains from about 1% to over 60%, but the per-model values appear only in a figure 2312.15524 §5, fig. 6. After fine-tuning on the survey, the best unblinded run reaches 0.113, against 0.134 for the best blinded run 2312.15524 tab. 1. Adding 10,000 purchase-only observations from Amazon browsing data worsened blinded evaluation (0.233) but left unblinded evaluation near its earlier error (0.120) 2312.15524 tab. 2.

The study covers one demand setting and one main model for the covariate analysis, and the human price range is not stated 2312.15524 §2, §3. All results are author-reported.

Step-level intervention for agent failures (CAR)

This variant applies the interventional question to a run of an LLM agent, not to a model's predictions. It asks which step of a failed run caused the failure 2606.08275 §1. The paper argues that the step which performs the harmful action is often only carrying out an earlier decision, so blaming that step is a category error 2606.08275 §1. It describes LLM-judge attribution as correlational, and cites about 14% step-level accuracy as the best reported result for it on the Who&When benchmark 2606.08275 §1 2606.08275 §6.

The intervention redraws the action at one step from the same policy and then reruns the rest of the run. The paper defines this resample as the null intervention. Four further operations are defined: forcing an action, replacing a tool result, editing the message history, and swapping the model from that step on 2606.08275 §2. Because the policy is stochastic, K reruns give a distribution of outcomes, not one path 2606.08275 §2. The single-step estimator compares the bad-outcome probability under the intervention with the observed run, and reports a confidence interval. The causal locus is the latest step whose interval still excludes zero on the side where the intervention lowers that probability 2606.08275 §4.

When two steps cause the failure only together, the paper splits credit with a Monte-Carlo Shapley estimator over steps. A coalition holds its steps at their factual actions, and its value is the bad-outcome probability over K rollouts. Marginals are not cached across permutations, because caching would make their variance collapse and give false confidence 2606.08275 §4.

The paper validates on synthetic structural causal models whose answers are planted in advance. There the single-step method recovers the pivotal step, and the Shapley estimator splits a two-step AND interaction at 0.44 and 0.45, against an analytic 0.5 each 2606.08275 §5. It reports no run on real agent traces and no comparison with LLM-judge or other attribution methods 2606.08275 §5 2606.08275 §7.

Two caveats follow from the design. The estimate is a total effect: rerunning step k also redraws every later random step, so an early and irrelevant step can look important. The paper names common random numbers as a refinement that it does not implement 2606.08275 §4 2606.08275 §7. The method also depends on the recorded state being faithful. Reruns call the environment afresh, so the side effects of real tools are not modelled, and the demonstrations use mocked tools 2606.08275 §7. Hosted models vary even at temperature 0, so the paper measures the rate at which replayed actions match the recorded ones, rather than assuming exact replay 2606.08275 §3.

When to use which

The review reports no benchmark that compares its counterfactual methods on speed or accuracy, so the choices below follow the mechanisms the review describes. Rows from the causal ML survey name the constraint the survey ties to each method. The survey also reports no head-to-head numbers.

Implementation sketch

The shared core is the instance-centric search. The review gives no learning rate, step count, λ schedule or start point for it, so those values are left open below 2103.04244 §6.1.

# Instance-centric counterfactual search  [§4.1] [§6.1]
# Input:  instance x (p,), model f giving a score, target y'
# Output: counterfactual x' (p,)

def counterfactual(x, target, model, lam, mad, steps):
    # Start point not stated; the instance is used here  [§6.1]
    candidate = x.copy()
    for _ in range(steps):                  # step count not stated  [§6.1]
        # Squared gap between the model score and the target  [eq. 10]
        gap = (model(candidate) - target) ** 2
        # L1 distance, each feature scaled by its MAD  [eq. 11]
        dist = sum(abs(candidate - x) / mad)
        # Trade the two terms off with lambda  [eq. 10]
        loss = lam * gap + dist
        # Optimiser is Adam; learning rate not stated  [§6.1]
        candidate = adam_step(candidate, grad(loss))
    return candidate

Variants differ in the loss (prototype, FACE path cost, DiCE diversity), in the optimiser (FISTA, Nelder-Mead, graph search) and in the distance 2103.04244 §6.1.

For an interventional or counterfactual query on a known SCM, the sketch above does not apply. Use the three-step procedure in the Formal core section instead 2206.15475 §2.3.2.

Contradictions

The sources frame causal grounding differently. The counterfactual review says post-hoc, model-agnostic explainers capture feature correlations rather than the mechanism behind a prediction, so they may not give causability 2103.04244 §1.2 2103.04244 §9.1. The causal ML survey places counterfactual explanations under causal explanations. It presents the smallest feature change that flips a prediction as one such explanation, and treats causal algorithmic recourse as the version that uses an SCM 2206.15475 §5.2.1 2206.15475 §5.2.2. These two papers disagree on whether the first kind is causal. Both sides are stated here and are not resolved.

The position paper argues that a symbolic description extracted from a network supports interventions and counterfactual questions 2012.05876 §5. This is the same claim as the counterfactual review's causal-grounding argument, from a different direction, and it is also not resolved. The position paper's causal claims are argued, not measured 2012.05876 §2.

The counterfactual review's own counts do not agree with each other. It says 26 algorithms were analysed in §5 and 18 in §6, and its summary table has 25 rows 2103.04244 §5 2103.04244 §6 2103.04244 tab. 2.

The review also reports a study by Zheng et al. with mixed results: causal information helped only users without prior domain knowledge [arxiv:2103.04244 §8.1] 2103.04244 §9.2.

The GPT-3 study's headline says the model fails the causal task 2206.14576 §Discussion. Its own table reports that intervention answers match the normative answer when B is present in the common-cause condition 2206.14576 fig. 5B. Both statements come from the same paper, and this page does not resolve them.

Papers

• Counterfactuals and Causability in Explainable Artificial Intelligence: Theory, Algorithms, and Applications: a systematic review of model-agnostic counterfactual explanations, with a seven-family taxonomy and an argument that they lack causal grounding.
• Causal Inference in Natural Language Processing: Estimation, Prediction, Interpretation and Beyond: a survey of causal inference for text, covering text as confounder, outcome or treatment, and causal views of robustness, fairness and interpretation. It runs no experiments.
• Causal Machine Learning: A Survey and Open Problems: a survey of causal ML, with the SCM core, five problem areas, a benchmark list and open problems. It runs no experiments.
• Neurosymbolic AI: The 3rd Wave: a position paper on combining neural learning with symbolic reasoning, with a view that extracted rules support interventional and counterfactual questions. It runs no experiments.
• General Pitfalls of Model-Agnostic Interpretation Methods for Machine Learning Models: a review of ten pitfalls in PDP, ALE, permutation importance and SHAP, with a section on why their output is not causal without assumptions.
• Using cognitive psychology to understand GPT-3: classic cognitive-psychology tests run on GPT-3, including a wine-cask task that compares intervention with observation. Results are author-reported.
• Multi-Agent Causal Discovery Using Large Language Models: a two-stage LLM debate method that picks and runs a statistical causal-discovery algorithm, then refines the graph with metadata. Results are author-reported.
• Tutor CoPilot: A Human-AI Approach for Scaling Real-Time Expertise: a randomised trial of an LLM tutoring assistant, with an intent-to-treat analysis and a 2SLS estimate of use. Results are author-reported.
• Inductive Biases for Deep Learning of Higher-Level Cognition: a position paper that proposes causal, verbalisable variables and sparse, local mechanism changes as inductive biases. It gives no algorithm or experiments.
• Using LLMs as simulated subjects, read through a causal lens: a study of LLM subjects in a price experiment, which shows confounding by the price and proposes disclosing the randomisation in the prompt. Results are author-reported.
• Causal Agent Replay: Counterfactual Attribution for LLM-Agent Failures: attributes a failed agent run to a step by resampling that step and rerunning, with a Monte-Carlo Shapley estimator for steps that interact. Validated only on synthetic models with planted answers.