Evaluating Large Language Models for Causal Modeling

Georg Niess; Houssam Razouk; Leonie Benischke; Roman Kern

REVIEW 3 major objections 3 minor 17 references

Evaluating Large Language Models for Causal Modeling

T0 review · 3 major / 3 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Large language models can assist human causal modelers, but the best model depends on the task: GPT-4-turbo and Llama3-70b lead at distilling text into causal variables, while Mixtral-8×22b is best at detecting interaction entities.

desk verdict Novel tasks, under-supported numbers: the labels come from the same LLMs being evaluated. read the letter →

arxiv 2411.15888 v1 pith:MBP5TE2F submitted 2024-11-24 cs.CL

Houssam Razouk , Leonie Benischke , Georg Niess , Roman Kern This is my paper · ORCID

classification cs.CL

keywords causalmodelinglargelanguagemodelsvariablesinteractionentitiesCohen'skappazero-shotevaluationdomaindependencydatascience

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper asks whether off-the-shelf large language models can do the early, human-heavy work of causal modeling: turning free-text domain knowledge into variables and spotting entities that combine several variables at once. It introduces two binary tasks, evaluates seven LLMs across eight domains in a zero-shot setting, and reports that LLMs are helpful but far from conclusive. The best agreement is a Cohen's $\kappa$ of about 0.40 for grouping entities under one causal variable and about 0.60 for detecting interaction entities, with task-specific winners: GPT-4-turbo and Llama3-70b lead the first task, while the sparse-expert Mixtral-8$\times$22b leads the second. If the results hold, LLMs could accelerate and broaden human expert causal modeling rather than replace it.

What carries the argument

The load-bearing objects are two binary classification tasks defined in the paper. Task 1, distilling causal domain knowledge into causal variables, asks whether two text entities are different values of the same causal variable (e.g., 'Insomniac' and 'Sound sleeper' both belong to 'Sleep quality'). Task 2, detecting interaction entities, asks whether one text entity represents values of two or more given causal variables simultaneously (e.g., 'Diabetic diet plan' spans 'Disease management' and 'Nutritional strategy'). Each LLM generates the dataset for eight domains, positive and negative examples are sampled, and every LLM classifies examples from its own and other models' data in a zero-shot setting with temperature zero; performance is scored by precision, recall, F1, and Cohen's $\kappa$. A cosine-similarity threshold classifier over text embeddings serves as a non-LLM reference to see whether the tasks reduce to semantic similarity.

What would settle it

Have two or more human domain experts independently annotate a held-out sample of entity pairs (Task 1) and interaction-entity triples (Task 2) across the same eight domains, then run the same zero-shot prompts and compute Cohen's $\kappa$ against the human labels; if the absolute values or the model ranking change materially, the paper's conclusion that these LLMs are the helpful tools described here is not supported.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that contemporary LLMs already support the two fundamental modeling steps that precede causal reasoning: deciding that two textual entities are values of the same causal variable rather than separate variables, and recognizing an entity as an interaction of several causal variables that should be represented as an artificial variable. The evaluation yields a consistent pattern: modern dense models (GPT-4-turbo, Llama3-70b) perform better at variable distillation, while sparse-expert Mixtral-8$\times$22b is best at interaction detection, and performance depends strongly on domain, with Health being the easiest for most models. The paper explicitly positions these findings as evidence that LLMs can be used in collaboration with human experts, providing a wider perspective, rather than as standalone causal modelers.

Load-bearing premise

The evaluation treats the positive and negative examples sampled from LLM-generated data as ground-truth labels for both tasks; if those generated groupings contain systematic errors, the measured kappa values and model rankings reflect agreement with machine-generated labels rather than true causal modeling ability.

Editorial extensions

If this is right

Off-the-shelf LLMs can already assist human experts in converting textual domain knowledge into causal variables and in flagging interaction entities, though agreement is only fair to moderate ($\kappa$ up to 0.40 and 0.60).
Model choice should be task-dependent: GPT-4-turbo and Llama3-70b are the strongest at variable distillation, while Mixtral-8$\times$22b is the strongest at interaction detection.
Domain matters more than raw model size for these tasks: most models score highest on Health data and much lower on Physics, suggesting a reliance on training-domain knowledge.
Simple semantic-similarity baselines do not explain LLM performance, so these tasks require genuine modeling rather than embedding lookup.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Because the evaluation uses LLM-generated data as ground truth, the reported kappas are best read as self-consistency with machine-generated labels; a human-expert-annotated set could produce different absolute values and possibly different rankings.
A natural extension is to measure human-LLM collaboration directly: whether expert modelers who see LLM suggestions produce causal graphs faster or with higher inter-expert agreement than modelers working alone.
The two tasks could form the core of a causal-modeling benchmark for future LLMs, but only after adding human-validated labels and non-English prompts.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Two things to know: the paper defines two genuinely new evaluation tasks for LLM-assisted causal modeling, and its headline rankings are not supported because the labels come from the same models being tested. The tasks are the real contribution; the numbers are a proof of concept.

What is new: Task 1 asks whether two text entities are values of the same causal variable; Task 2 asks whether an entity is an interaction of given causal variables. Both are natural and previously untested steps in building causal models from text, and the paper grounds them in the causal modeling literature (Suzuki, McDonnell, Kuorikoski). The experimental design is coherent: seven models, eight domains, zero-shot, deterministic sampling, and the authors include a cosine-similarity analysis to probe when predictions track surface semantics. Code and data are released, which helps.

The soft spot is the one the authors admit in the Limitations: 'these exact numbers have been derived based on generated data.' Every model generates the variables and interaction entities, those generations are sampled into positives and negatives, and the same models then classify them. There is no expert-validated gold standard. So precision, recall, and kappa measure agreement with machine-generated labels, not correctness against a known causal structure. The ranking of GPT-4-turbo/Llama3 on Task 1 and Mixtral on Task 2 could just reflect self-consistency. The cosine-similarity plots reinforce this: several models reach peak agreement at high similarity thresholds, suggesting they are doing surface matching. The paper also reports single numbers without confidence intervals or significance tests, so the differences between models are not statistically grounded.

I would not call this a fatal flaw; the framing is transparent and the tasks are worth keeping. But the central claim that some LLMs are 'helpful tools' for causal modeling, and the specific model rankings, need external validation before they can be used. A serious referee should ask for a human/domain-expert labeled subset for each domain, at least a few hundred items, plus variance across repeated sampling.

Who is this for: people building LLM-assisted pipelines for causal modeling, and benchmark designers. The paper deserves a round of peer review. Desk rejection would be wrong, because the tasks are novel and the authors are honest about the limitation. Send it to review with the expectation that the empirical section gets reworked.

Referee Report

3 major / 3 minor

Summary. The paper introduces two tasks for evaluating LLMs' ability to support causal modeling: (Task 1) deciding whether two textual entities are values of the same causal variable, and (Task 2) deciding whether a textual entity represents an interaction of multiple given causal variables. For each task, seven LLMs generate candidate data (causal variables, values, interaction entities) in several domains, the generated data is sampled into positive and negative examples, and all LLMs classify all examples. Performance is measured by precision, recall, F1, and Cohen's kappa, with additional cosine-similarity analysis. The authors claim that GPT-4-turbo and Llama3-70b perform best on Task 1, while Mixtral-8×22b performs best on Task 2, and conclude that LLMs are helpful collaborators for human causal modelers.

Significance. If the central claims were established, the paper would provide a useful step toward benchmarking LLMs for a relatively underexplored component of causal analysis: transforming domain knowledge into causal variables and interaction entities. The task formulations themselves are interesting and connect to established guidelines from causal data science (Suzuki et al., McDonnell, Kuorikoski). However, the empirical evaluation is built on machine-generated labels with no expert-validated gold standard, so the numerical findings and model rankings are not currently supported. The paper does transparently acknowledge this limitation, which is a point in its favor, but the acknowledgement does not repair the circularity.

major comments (3)

[Section 3.2 and Limitations] The evaluation is circular: the ground-truth labels for both tasks are derived from data generated by the very LLMs being evaluated. Section 3.2 (Figure 2) states that each LLM generates causal variables, values, and interaction entities, and these outputs are sampled into positive and negative examples; the same LLMs then classify those examples. The Limitations section concedes that 'these exact numbers have been derived based on generated data' and that the findings need to be compared with domain-specific experts. Consequently, the precision, recall, and kappa values in Tables 1-4 measure agreement with a machine-generated labeling scheme, not correctness with respect to true causal variable structure. The abstract's claim that 'contemporary LLMs are helpful tools' for causal modeling, and the specific rankings (e.g., Table 1: GPT-4-turbo and Llama3-70b with kappa 40% and 39%; Table 2: Mixtral-8×22b with kappa 60%), are not established by the presented evidence.
[Tables 1-4] The reported results are single-point estimates with no measures of uncertainty. Each kappa, precision, and recall value appears to be computed once over a pooled dataset, yet the experiment includes multiple domains and multiple generating models. There is no confidence interval, bootstrap, or significance test. For example, in Table 2, Mixtral-8×22b (kappa 60%) and Llama3-70b (kappa 47%) differ by 13 percentage points, but without variance information it is impossible to know whether this difference is meaningful. The paper should either provide per-domain and per-generator breakdowns with uncertainty or refrain from claiming that one model 'stands out' over others.
[Figures 3-6 and Section 4] The cosine-similarity analyses raise a specific confound that is acknowledged but not resolved. Figures 4 and 6 show that several models' predictions align most strongly with classifying entities as belonging to the same variable (Task 1) or as interaction entities (Task 2) when their embedding cosine similarity is high. This suggests that models may be exploiting surface similarity rather than causal semantics. The paper notes this trend for weaker models but does not control for it in the main kappa-based comparison. Without such a control, the superior kappa values of the top performers could reflect better alignment with the (machine-generated) labels rather than genuine causal modeling ability.

minor comments (3)

[Section 4] The list of domains in Section 4 includes 'Sociology' but the preceding sentence lists '(i) Computer Science ... and (vii) Urban Studies'; the count and the labels are inconsistent. Also, Table 3 and Table 4 omit columns for GPT-3.5-turbo, Mixtral-8×7b, and Mistral-7b without explanation, even though these models are part of the main results.
[Abstract and affiliation] The author affiliation on the title page reads 'Université Bourgogne Franche-Comté, Besançon, Franc' — the final word should be 'France'.
[Section 5] The Discussion interprets the higher precision than recall as potentially caused by Temperature=0, but this is speculative and not supported by any comparison with nonzero temperature. A sentence acknowledging this as an untested hypothesis would be more accurate.

Circularity Check

1 steps flagged · score 6.0 of 10

Central kappa-based rankings rest on LLM-generated labels, so the measured 'causal modeling ability' is self-consistency with machine-generated ontologies rather than expert-validated correctness.

self definitional [Section 3.2, Figure 2; Section 4 (Tables 1 and 2); Limitations]
"As depicted in Figure 2, our experiments use a within-subjects design involving all participating LLMs in every stage. Each LLM generates two datasets and evaluates its own generated data and that of other LLMs. ... Cohen's kappa coefficient [Cohen, 1960] is employed to estimate the agreement between the predictions of the LLM and the generated data of that LLM and other LLMs under study. ... Limitations: This study significantly narrow down the search space for the appropriate LLM for causal modeling. However, these exact numbers have been derived based on generated data."

For both Task 1 and Task 2, the positive and negative labels that serve as ground truth are constructed by sampling from entities generated by the same LLMs that are later evaluated. Precision, recall, and Cohen's kappa therefore quantify agreement with LLM-generated causal-variable groupings and interaction labels, not with an externally validated causal structure. A model that generates semantically coherent but non-expert ontologies and then classifies its own generated items will score high purely by self-consistency; cross-model evaluation does not remove this because every label still originates from an LLM. The paper's own Limitations section concedes that the reported numbers 'have been derived based on generated data' and require comparison with domain experts.

full rationale

The paper is transparent about its main limitation, and not every use of LLM-generated data is circular: the task definitions are grounded in external causal-modeling literature (Suzuki et al. 2020, McDonnell 2018), the evaluation is cross-model (each model scores data generated by other models as well as itself), and the reported kappas are not numerically forced—models can and do disagree with generated labels. However, the central empirical quantity—the kappa-based ranking of LLMs on Task 1 and Task 2—is measured against labels that were themselves produced by the model class under evaluation. No human expert gold standard or externally validated causal graph is used. This makes the headline finding a measure of agreement with machine-generated causal ontologies, not of correctness relative to expert causal modeling. The self-citation to Razouk and Kern (2022) is not load-bearing, and there is no imported uniqueness theorem or ansatz-by-citation. The circularity is therefore partial rather than total: the derivation is not an identity, but the main conclusion would need independent human-validated labels to escape the self-referential evaluation loop.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The paper introduces no new particles, forces, dimensions, or physical entities. Its contributions are evaluation tasks built on prior causal modeling constructs. The key ledger entry is the assumption that LLM-generated data can serve as ground truth, which is an ad hoc premise not validated against an external standard.

assumptions (3)

domain assumption Entities should be grouped under causal variables rather than kept as independent entities.
Task 1's correctness assumes the grouping recommendation from Suzuki et al. 2020 and related causal modeling guidelines cited in Sections 1 and 2.1.
domain assumption Each text entity represents at least one value of at least one causal variable.
Section 3.1 problem formulation: "Each entity e in E represents at least one value v_i,j of a causal variable V_i in V." This mapping makes the positive and negative labels well-defined.
ad hoc to paper LLM-generated samples are valid ground truth for evaluating LLMs.
Section 3.2 constructs labels exclusively from LLM-generated data, and the Limitations section acknowledges that expert comparison is still needed. This unverified premise underlies all quantitative claims.

how reviews work

0 comments

Cite this review

Pith. "Pith review of Evaluating Large Language Models for Causal Modeling." pith.science (2026). https://pith.science/paper/MBP5TE2F

@misc{pith2026241115888,
  author       = {Pith},
  title        = {Pith review of: Evaluating Large Language Models for Causal Modeling},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/MBP5TE2F}},
  note         = {Machine review of arXiv:2411.15888}
}

read the original abstract

In this paper, we consider the process of transforming causal domain knowledge into a representation that aligns more closely with guidelines from causal data science. To this end, we introduce two novel tasks related to distilling causal domain knowledge into causal variables and detecting interaction entities using LLMs. We have determined that contemporary LLMs are helpful tools for conducting causal modeling tasks in collaboration with human experts, as they can provide a wider perspective. Specifically, LLMs, such as GPT-4-turbo and Llama3-70b, perform better in distilling causal domain knowledge into causal variables compared to sparse expert models, such as Mixtral-8x22b. On the contrary, sparse expert models such as Mixtral-8x22b stand out as the most effective in identifying interaction entities. Finally, we highlight the dependency between the domain where the entities are generated and the performance of the chosen LLM for causal modeling.

Figures

Figures reproduced from arXiv: 2411.15888 by the authors.

**Figure 1.** Causal modeling of domain knowledge by human experts. This is a tedious and error prone process. We study the usefulness of LLMs to support this process, whereby we first divide this into two distinct tasks. drawbacks, including violating the modality of the causal model in the variables [Kuorikoski, 2012, Suzuki et al., 2020]. Also, considering the interaction entity "Diabetic diet plan", this entity describes the … view at source ↗

**Figure 2.** Experiments workflow on benchmarking LLMs causal modeling abilities. Each LLM is instructed to generate data sets according to the task and the domain. The generated data sets are sampled to construct positive and negative examples. Each LLM is instructed to evaluate positive and negative examples sampled from its own and other LLMs generated data in a zero-shot setting. Experiment on Task 1: Distilling Causal Domai… view at source ↗

**Figure 3.** The agreement between the generated data for Task 1 and the classification results based on the cosine similarity threshold. Agreement at a lower cosine similarity threshold indicates that, in the generated data, two entities represent different values of the same causal variable and are not necessarily semantically similar. The data generated by Mixtral-8×22b tends to have a higher cosine similarity to other LLMs. … view at source ↗

Figures from the paper (4 more)

**Figure 4.** Figure 4: The agreement between LLMs predictions and the classification is based on the cosine similarity threshold. GPT-3.5-turbo, Mixtral-8×7b and Mistral-7b reach their peak agreement at a higher threshold, less aligned with the generated data. GPT-4-turbo, Llama3-70b and Lla…

**Figure 5.** Figure 5: The agreement between the generated data for Task 2 and the classification results based on the cosine similarity threshold. Agreement at a lower cosine similarity threshold indicates that a text represents values of different causal variables and is not necessarily se…

**Figure 6.** Figure 6: The agreement between the LLMs predictions on Task 2 and the classification results based on the cosine similarity threshold. Higher agreement with the model predicting at higher similarity thresholds indicates that LLMs, to a certain extent, predict a text as values o…

**Figure 4.** Figure 4: This trend is less dominant for Task 2 as shown in Figure 6. Finally, the study unveils a significant dependency [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

17 extracted references · 8 canonical work pages

[3]

Causal parrots: Large language models may talk causality but are not causal

Matej Zeˇcevi´c, Moritz Willig, Devendra Singh Dhami, and Kristian Kersting. Causal parrots: Large language models may talk causality but are not causal. arXiv preprint arXiv:2308.13067,

arXiv
[7]

Spock at fincausal 2022: Causal information extraction using span-based and sequence tagging models

Anik Saha, Jian Ni, Oktie Hassanzadeh, Alex Gittens, Kavitha Srinivas, and Bulent Yener. Spock at fincausal 2022: Causal information extraction using span-based and sequence tagging models. In Proceedings of the 4th Financial Narrative Processing Workshop@ LREC2022, pages 108–111,

work page 2022
[8]

Causal bert: Language models for causality detection between events expressed in text

Vivek Khetan, Roshni Ramnani, Mayuresh Anand, Shubhashis Sengupta, and Andrew E Fano. Causal bert: Language models for causality detection between events expressed in text. arXiv preprint arXiv:2012.05453,

arXiv 2012
[9]

Event causality extraction via implicit cause-effect interactions

Jintao Liu, Zequn Zhang, Kaiwen Wei, Zhi Guo, Xian Sun, Li Jin, and Xiaoyu Li. Event causality extraction via implicit cause-effect interactions. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 6792–6804,

work page 2023
[10]

Causal reasoning and large language models: Opening a new frontier for causality

Emre Kıcıman, Robert Ness, Amit Sharma, and Chenhao Tan. Causal reasoning and large language models: Opening a new frontier for causality. arXiv preprint arXiv:2305.00050,

arXiv
[11]

Lego: A multi-agent collaborative framework with role-playing and iterative feedback for causality explanation generation

Zhitao He, Pengfei Cao, Yubo Chen, Kang Liu, Ruopeng Li, Mengshu Sun, and Jun Zhao. Lego: A multi-agent collaborative framework with role-playing and iterative feedback for causality explanation generation. In Findings of the Association for Computational Linguistics: EMNLP 2023 , pages 9142–9163,

work page 2023
[12]

Causal-discovery performance of chatgpt in the context of neuropathic pain diagnosis, jan

12 A PREPRINT - NOVEMBER 26, 2024 R Tu, C Ma, and C Zhang. Causal-discovery performance of chatgpt in the context of neuropathic pain diagnosis, jan

work page 2024
[13]

The causal reasoning ability of open large language model: A comprehensive and exemplary functional testing

Shun-Hang Li, Gang Zhou, Zhi-Bo Li, Ji-Cang Lu, and Ning-Bo Huang. The causal reasoning ability of open large language model: A comprehensive and exemplary functional testing. In 2023 IEEE 23rd International Conference on Software Quality, Reliability, and Security (QRS) , pages 240–249. IEEE,

work page 2023

Show all 17 references

[14]

Gpt-4 technical report

Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. Gpt-4 technical report. arXiv preprint arXiv:2303.08774,

arXiv
[15]

Llama: Open and efficient foundation language models (2023)

Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timothée Lacroix, Baptiste Rozière, Naman Goyal, Eric Hambro, Faisal Azhar, et al. Llama: Open and efficient foundation language models (2023). arXiv preprint arXiv:2302.13971,

2023 arXiv
[16]

Mixtral of experts

Albert Q Jiang, Alexandre Sablayrolles, Antoine Roux, Arthur Mensch, Blanche Savary, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Emma Bou Hanna, Florian Bressand, et al. Mixtral of experts. arXiv preprint arXiv:2401.04088,

arXiv
[1960]

Richard Hahn, and Huan Liu

11 A PREPRINT - NOVEMBER 26, 2024 Ruocheng Guo, Lu Cheng, Jundong Li, P. Richard Hahn, and Huan Liu. A survey of learning causality with data. ACM Computing Surveys , 53(4):1–37, Sep

2024
[2009]

doi: 10.1214/08-aos613

ISSN 0090-5364. doi: 10.1214/08-aos613. URL http://dx.doi.org/10.1214/ 08-AOS613. Zhijing Jin, Jiarui Liu, Zhiheng Lyu, Spencer Poff, Mrinmaya Sachan, Rada Mihalcea, Mona Diab, and Bernhard Schölkopf. Can large language models infer causation from correlation? arXiv preprint a...

arXiv
[2020]

doi: 10.1145/3397269

ISSN 1557-7341. doi: 10.1145/3397269. URL http: //dx.doi.org/10.1145/3397269. Sander Greenland and Babette Brumback. An overview of relations among causal modelling methods. International journal of epidemiology, 31(5):1030–1037,

doi
[2022]

Tc-gat: Graph attention network for temporal causality discovery

Xiaosong Yuan, Ke Chen, Wanli Zuo, and Yijia Zhang. Tc-gat: Graph attention network for temporal causality discovery. arXiv preprint arXiv:2304.10706,

arXiv
[2023]

Weakly supervised multilingual causality extraction from wikipedia

Chikara Hashimoto. Weakly supervised multilingual causality extraction from wikipedia. In Proceedings of the 2019 conference on empirical methods in natural language processing and the 9th international joint conference on natural language processing (emnlp-ijcnlp), pages 2988–2999,

2019
[2024]

Mistral 7b

Albert Q Jiang, Alexandre Sablayrolles, Arthur Mensch, Chris Bamford, Devendra Singh Chaplot, Diego de las Casas, Florian Bressand, Gianna Lengyel, Guillaume Lample, Lucile Saulnier, et al. Mistral 7b. arXiv preprint arXiv:2310.06825,

arXiv

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.

[1] [3]

Causal parrots: Large language models may talk causality but are not causal

Matej Zeˇcevi´c, Moritz Willig, Devendra Singh Dhami, and Kristian Kersting. Causal parrots: Large language models may talk causality but are not causal. arXiv preprint arXiv:2308.13067,

arXiv

[2] [7]

Spock at fincausal 2022: Causal information extraction using span-based and sequence tagging models

Anik Saha, Jian Ni, Oktie Hassanzadeh, Alex Gittens, Kavitha Srinivas, and Bulent Yener. Spock at fincausal 2022: Causal information extraction using span-based and sequence tagging models. In Proceedings of the 4th Financial Narrative Processing Workshop@ LREC2022, pages 108–111,

work page 2022

[3] [8]

Causal bert: Language models for causality detection between events expressed in text

Vivek Khetan, Roshni Ramnani, Mayuresh Anand, Shubhashis Sengupta, and Andrew E Fano. Causal bert: Language models for causality detection between events expressed in text. arXiv preprint arXiv:2012.05453,

arXiv 2012

[4] [9]

Event causality extraction via implicit cause-effect interactions

Jintao Liu, Zequn Zhang, Kaiwen Wei, Zhi Guo, Xian Sun, Li Jin, and Xiaoyu Li. Event causality extraction via implicit cause-effect interactions. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 6792–6804,

work page 2023

[5] [10]

Causal reasoning and large language models: Opening a new frontier for causality

Emre Kıcıman, Robert Ness, Amit Sharma, and Chenhao Tan. Causal reasoning and large language models: Opening a new frontier for causality. arXiv preprint arXiv:2305.00050,

arXiv

[6] [11]

Lego: A multi-agent collaborative framework with role-playing and iterative feedback for causality explanation generation

Zhitao He, Pengfei Cao, Yubo Chen, Kang Liu, Ruopeng Li, Mengshu Sun, and Jun Zhao. Lego: A multi-agent collaborative framework with role-playing and iterative feedback for causality explanation generation. In Findings of the Association for Computational Linguistics: EMNLP 2023 , pages 9142–9163,

work page 2023

[7] [12]

Causal-discovery performance of chatgpt in the context of neuropathic pain diagnosis, jan

12 A PREPRINT - NOVEMBER 26, 2024 R Tu, C Ma, and C Zhang. Causal-discovery performance of chatgpt in the context of neuropathic pain diagnosis, jan

work page 2024

[8] [13]

The causal reasoning ability of open large language model: A comprehensive and exemplary functional testing

Shun-Hang Li, Gang Zhou, Zhi-Bo Li, Ji-Cang Lu, and Ning-Bo Huang. The causal reasoning ability of open large language model: A comprehensive and exemplary functional testing. In 2023 IEEE 23rd International Conference on Software Quality, Reliability, and Security (QRS) , pages 240–249. IEEE,

work page 2023