Pith. sign in

REVIEW 3 major objections 5 minor 67 references

Mind the XAI Gap: A Human-Centered LLM Framework for Democratizing Explainable AI

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

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

Pith's one-line read A structured LLM prompt with a few LIME examples produces expert-grade feature rankings and clear explanations for non-experts in one response.

desk verdict Worth engaging: a genuinely new LLM+XAI framework with a real user study, but the headline Spearman=0.92 is weaker than it looks because LIME both supplies the in-context examples and serves as the scoring target. read the letter →

arxiv 2506.12240 v1 pith:7UPYDI6Z submitted 2025-06-13 cs.LG cs.AI

Eva Paraschou , Ioannis Arapakis , Sofia Yfantidou , Sebastian Macaluso , Athena Vakali This is my paper
classification cs.LGcs.AI
keywords ExplainableAI(XAI)LargeLanguageModelsHuman-centeredexplainabilityIn-contextlearningFeatureimportanceUserstudyWell-beingclusteringLIME
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

The paper claims that a carefully structured LLM prompt, fed with a handful of LIME-style explanation examples, can produce both expert-grade feature-importance rankings and plain-language explanations for non-experts in a single response. The framework is built around a ground-truth contextual 'thesaurus' derived from a well-being clustering analysis, and it is meant to close three gaps in current XAI: black-box transparency, explanations that make no sense to non-experts, and the costly post-processing step needed to humanize raw XAI output. If the claim holds, high-stakes domains like healthcare could skip the standard three-step pipeline of model inference, explanation extraction, and manual interpretation, and instead hand both audiences one immediately usable explanation. The paper reports a Spearman rank correlation of 0.92 with LIME feature rankings (LLaMA3, few-shot) and a user study (N=56) in which the framework's explanations scored significantly higher than LIME on pragmatic quality.

What carries the argument

The load-bearing mechanism is the ground-truth contextual 'thesaurus': a benchmark set of well-being cluster assignments and LIME feature-importance explanations, built from over 40 data, model, and XAI combinations, which is fed into the LLM through structured prompt and system templates. The prompt supports zero-, one-, and few-shot modes; the few-shot mode inserts several demonstration examples of the form output-features-explanation. The system template has three parts: 'Setting the Scene' tells the LLM to compute feature importance; 'Expert Response' forces a JSON list of features with importance values; 'Non-expert Response' turns those values into a plain-language paragraph. Everything hinges on in-context learning transferring the LIME-style ranking procedure from the demonstrations to new instances, which is what moves Spearman correlation from near zero in zero-shot to 0.92 in few-shot.

What would settle it

Replace the few-shot demonstrations with a deliberately permuted LIME ranking for the same instances while keeping the prompt otherwise identical; if the LLM's output correlates with the permuted ranking instead of the original LIME ranking, then the reported Spearman=0.92 measures imitation of the demonstrations, not correctness of the explanation.

Watch

Extended reading notes

Core claim

The central claim is that LLMs, when given in-context demonstrations of a foundational XAI method, can act as a full human-centered explainability layer: they reproduce the technical feature-importance rankings an expert would get from LIME, and simultaneously translate those rankings into a human-friendly paragraph a non-expert can act on. The paper constructs the demonstrations by clustering the LifeSnaps well-being data with k-means, converting the cluster assignments into a classification problem (99% accuracy with a support vector classifier), and generating LIME explanations on those predictions; the best clusterings and LIME explanations form the thesaurus that is injected into the prompt. In the few-shot setting LLaMA3 reaches Spearman rank correlation 0.92 with LIME's rankings, with NDCG difference 0.001 and Euclidean distance 0.02, while Mistral reaches 0.74, and zero-shot performance is close to zero, showing that the contextual demonstrations, not the LLM's prior knowledge, carry the content quality. In a paired user study, the framework's human-centered explanations outperformed LIME on overall pragmatic quality (1.00 vs -0.16) and on every pragmatic item, with hedonic gains present but not statistically significant. The paper positions this as evidence that LLMs are viable enablers of human-centered explainable AI.

Load-bearing premise

The load-bearing premise is that LIME's feature-importance rankings are the correct explanation, and because those same rankings are also the few-shot demonstrations, the measured correlation measures imitation of the examples rather than external correctness.

Editorial extensions

If this is right

  • Few-shot demonstrations are what make the framework work: LLaMA3's rank correlation with LIME jumps from 0.01 zero-shot to 0.92 few-shot, so the contextual examples, not the pre-trained model alone, carry the explanation quality.
  • The framework collapses the usual three-step pipeline (model inference, explanation extraction, post-processing) into a single LLM response for new instances.
  • For non-experts, the framework's explanations beat raw LIME on user-rated clarity, ease of use, efficiency, and supportiveness, with the paired overall pragmatic-quality score rising from -0.16 to 1.00.
  • Because the prompt and system templates are designed to be data-, model-, and explanation-agnostic, the same scaffolding can be rebuilt for other domains and tasks by supplying a new thesaurus.
  • Content quality is measured against LIME, so the framework's 'high-quality technical explanation' claim is bounded by LIME's fidelity (0.93 on the benchmarking sample) as the reference standard.

Reading between the lines

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

  • A natural test of whether the LLM is computing feature importance or imitating its examples is to swap the few-shot demonstrations to a different XAI method (e.g., SHAP or Anchor): if the outputs track the new method, the framework is a translation layer for whatever explanation it is shown rather than an independent reasoner.
  • The user-study gain may partly reflect modality rather than LLM reasoning, since the comparison is text-plus-JSON against LIME's raw visualization; a human-written plain-language translation of the same LIME output would isolate how much the LLM adds.
  • The agnosticism claim could be stress-tested by rebuilding the thesaurus in a different high-stakes domain, such as credit decisions or clinical risk scores, and checking whether the Spearman correlation and user preference survive the transfer; degradation would show the thesaurus, not the templates, is the bottleneck.
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.

Referee Report

3 major / 5 minor

Summary. The paper proposes an LLM-based framework for human-centered explainable AI (HCXAI) that uses in-context learning to generate, in a single response, both technical explanations for experts and human-friendly explanations for non-experts. The framework is instantiated on a well-being clustering task using the LifeSnaps dataset, with a ``contextual thesaurus'' built from LIME explanations over 20 instances. The authors evaluate the framework through a user study (N=56) comparing their explanations with LIME, and through automatic structure and content quality metrics on 20 instances. They report a Spearman rank correlation of 0.92 with LIME ground truth and significant improvements in pragmatic quality over LIME in a paired comparison, concluding that LLMs can serve as HCXAI enablers.

Significance. If the claims were fully supported, the work would be a useful contribution to the growing literature on using LLMs to make XAI accessible, and the reproducible code release and detailed comparison of prompting techniques are commendable. The user study provides real, if partial, evidence that the framework's explanations are preferred by non-experts over raw LIME output in a paired comparison, particularly on pragmatic quality. However, the central quantitative claim of high content quality (Spearman=0.92, Abstract and Section 5.2) is undermined by a circular evaluation setup: the same LIME explanations that serve as few-shot demonstrations in the prompt are also used as the scoring ground truth. The reported numbers therefore measure how well the LLM imitates its in-context examples, not whether the explanations are correct in any external sense. The paper is well-structured and transparent about its methodology, which makes the circularity easy to identify, but the main evidence for technical content quality needs substantial rework before the central claim can be accepted.

major comments (3)
  1. [Sections 3.2, 4, and 5.2, Table 4] The Spearman=0.92 result is not a valid measure of content quality because of circularity. In Section 3.2, LIME explanations are selected as the ground-truth ``thesaurus''; in Section 4, the few-shot prompt includes those same LIME explanations as demonstration examples; and in Section 5.2, the framework's outputs are compared against LIME on 20 randomly selected instances. The paper does not state whether the 20 evaluation instances are disjoint from the few-shot demonstration instances. If they are not disjoint, the comparison reduces to testing how well the LLM copies its input examples. Even if they are disjoint, the zero-to-few-shot trajectory in Table 4 (LLaMA3 Spearman rising from 0.01 to 0.92) is exactly what would be expected from in-context imitation of the output format and feature-name style, and does not establish correctness with respect to any external notion of feature importance. The authors should either (a) evaluate against a ground truth that is independent of the demonstrations (e.g., human expert feature judgments on the same instances), or (b) at minimum demonstrate that the evaluation instances are disjoint from the demonstrations and that the correlation persists when the demonstrations are replaced by LIME explanations on different instances, and clearly report this separation.
  2. [Section 3.2, Table 1] The designation of LIME as ground truth rests on a fidelity of 0.93, but LIME fidelity measures how well the local surrogate agrees with the classifier's predictions, not whether the identified features are causally responsible for the well-being clusters. The authors reformulate clustering as a classification task (with 99% accuracy) and then apply LIME to the classifier, so any errors in the surrogate or in the reformulation are inherited by both the demonstrations and the evaluation metric. This is a correctness-risk concern, not just a circularity issue: if LIME is inaccurate for this clustering problem, the ``high content quality'' claim collapses regardless of the Spearman value. The authors should provide an independent validation of the feature importance rankings, or explicitly temper the claim from ``high content quality'' to ``high similarity to LIME'' and discuss what that does and does not establish.
  3. [Section 5.1, Tables 2 and 3] The summary statement that the framework ``significantly outperforms LIME in a user study'' is stronger than the results support. In the A/B test (Table 2), the confidence intervals all include zero for every item and overall scale, and the authors correctly note the absence of statistically significant differences there. The significant advantage appears only in the paired comparison (Table 3), where it is confined to Pragmatic Quality and its four items; the Hedonic Quality overall difference (0.11 vs. -0.27) is not statistically significant, and the confidence intervals for several HQ items include zero. The abstract and conclusion should be reworded to state that the paired comparison showed significant improvements in pragmatic quality, while hedonic quality improvements were not significant and the independent A/B test showed only non-significant trends. This does not invalidate the user-study evidence, but it narrows the claim considerably.
minor comments (5)
  1. [Abstract and Section 1] The phrase ``Alarming evidence about AI's unprecedented human values risks'' is grammatically unclear; consider revising to ``risks to human values'' or similar.
  2. [Section 4, Figure 5] The colors (light blue, orange, light green) used to mark configurable parts in the prompt and system templates are not visually annotated in the text, making it hard to follow which parts are data, model, and explanation-specific. Please add a legend or describe the color coding more explicitly in the caption.
  3. [Section 5.1, Tables 2 and 3] The tables are dense; adding a note that confidence intervals are 95% confidence intervals and stating the sample size per group in the caption (N=25 for LIME, N=31 for framework in Table 2) would improve interpretability.
  4. [Section 5.2, Table 4] The row labels ``structure'' and ``content'' are ambiguous; the table header reads ``structure Content''. Clarify that structure quality comprises coherence, grammar errors, readability, and sentiment consistency, while content quality comprises Spearman, NDCG difference, and Euclidean distance.
  5. [Section 6, Limitations] The limitation that ``the current implementation supports only realistic features'' is valuable, but the discussion of hallucinations equates low NDCG difference with reduced hallucination; this is an indirect proxy and should be stated as such rather than as a direct measure of factual accuracy.

Circularity Check

1 steps flagged · score 6.0 of 10

Content-quality claim is circular: LIME supplies the few-shot demonstrations (Section 4) and is then used as the 'ground-truth' benchmark (Section 5.2), so Spearman=0.92 measures in-context imitation rather than correctness.

  1. fitted input called prediction [Sections 3.2, 4, and 5.2; Table 4]
    "Our “thesaurus” provides rich domain- (i.e. Lifesnaps multi-modal data, well-being clusters) and explainability-relevant (e.g., raw LIME explanations) knowledge, which will be transferred to the LLM using in-context learning prompting techniques ... The content quality evaluation quantifies the similarity between explanations generated by the ground-truth XAI method (i.e., LIME) and those generated by our framework, and thus, we use the following metrics: Spearman rank correlation ..."

    The same LIME explanations serve as the few-shot demonstrations in Section 4 and as the 'ground-truth' benchmark in Section 5.2. Spearman rank correlation therefore measures how closely the LLM reproduces its own input examples, not whether the explanations are correct by any external standard. Table 4 makes the mechanism explicit: LLaMA3's Spearman rises from 0.01 in zero-shot (no LIME examples) to 0.92 in few-shot (LIME examples in the prompt), which is the expected signature of in-context imitation. The paper's own limitations acknowledge that factual correctness was not assessed, listing 'assessing factual errors in explanations' as future work, so the headline claim of high content quality reduces to agreement with the input demonstrations.

full rationale

The central quantitative evidence for content quality is circular. In Section 3.2 LIME is selected as the ground-truth explainability source based on its 0.93 fidelity, and its instance-level feature importance forms the 'thesaurus.' Section 4 transfers this thesaurus into the prompt via one/few-shot demonstrations, explicitly including raw LIME explanations. Section 5.2 then scores the framework by Spearman rank correlation between its feature rankings and LIME's rankings. Table 4's trajectory, with LLaMA3 Spearman rising from 0.01 in zero-shot to 0.92 in few-shot, is exactly what in-context imitation predicts: with no demonstrations the model lacks format and content to copy, and with LIME examples it reproduces LIME rankings. Thus the headline 'high content quality' claim in the Abstract and Section 5.2 is not an external validation; it measures agreement with the input examples. The paper's own limitations acknowledge this gap by listing 'assessing factual errors in explanations' as future work. The user study in Section 5.1 is independent evidence for human-friendliness in a paired comparison, so the framework does have non-circular support for the human-centered part; however, the technical content-quality claim, central to the contribution of providing 'high-quality technical explanations aligned with foundational XAI methods,' reduces to a fit. Score 6 reflects this partial circularity, with the independent user study preventing a higher score.

Assumptions & free parameters 4 free parameters · 4 assumptions · 1 invented entities

The framework's central claim rests on several hand-picked choices: the number of clusters, the selection of LIME as the explanation source, and unspecified few-shot count and decoding parameters. The most consequential is the use of LIME as both the demonstration material and the evaluation ground truth, which creates circularity in the content-quality assessment.

free parameters (4)
  • Number of clusters (k) = 2
    Chosen by silhouette score after elbow method and outlier removal; affects cluster labels and all downstream explanations.
  • LIME selection as ground truth = fidelity 0.93
    LIME was chosen as the explainability ground truth after observing its high fidelity among four methods (Table 1), a post hoc selection that the evaluation then treats as ground truth.
  • Number of few-shot demonstration examples = not specified
    The paper says 'few-shot' but does not report the exact number of demonstrations used, which is a key prompt parameter affecting content quality.
  • LLM decoding temperature = not reported
    Decoding parameters are not given, so explanation variability is uncontrolled.
assumptions (4)
  • domain assumption LIME explanations on a surrogate classifier reflect the true feature attributions of the original k-means clustering.
    The paper reformulates clustering as classification with an SVC (99% accuracy) and applies LIME to the surrogate, assuming this transfers to the clustering model. Introduced in Section 3.2.
  • domain assumption LLMs can accurately compute feature importance from feature values and cluster labels when given in-context examples.
    The framework relies on the pre-trained and in-context abilities of Mistral and LLaMA to output importance values; the paper acknowledges in Section 6 that zero-shot performance is poor.
  • domain assumption The well-being cluster labels derived from the validation set are meaningful for feature importance.
    The two k-means clusters are labeled as positive/negative well-being using Mann-Whitney tests on validation features; the labels are then used as class targets in the surrogate classifier.
  • domain assumption The User Experience Questionnaire is a valid instrument for measuring interpretability and human-friendliness of explanations.
    The study uses UEQ short scales; the paper does not validate the instrument for this specific task.
invented entities (1)
  • Ground-truth contextual thesaurus
    purpose: A benchmark-base of LIME explanations and cluster information used as few-shot demonstrations to guide LLMs.
    The thesaurus is a constructed artifact whose 'ground-truth' status is inherited from LIME, not independently verified; its validity is the main circularity issue.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Mind the XAI Gap: A Human-Centered LLM Framework for Democratizing Explainable AI." pith.science (2026). https://pith.science/paper/7UPYDI6Z

@misc{pith2026250612240,
  author       = {Pith},
  title        = {Pith review of: Mind the XAI Gap: A Human-Centered LLM Framework for Democratizing Explainable AI},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7UPYDI6Z}},
  note         = {Machine review of arXiv:2506.12240}
}
read the original abstract

Artificial Intelligence (AI) is rapidly embedded in critical decision-making systems, however their foundational ``black-box'' models require eXplainable AI (XAI) solutions to enhance transparency, which are mostly oriented to experts, making no sense to non-experts. Alarming evidence about AI's unprecedented human values risks brings forward the imperative need for transparent human-centered XAI solutions. In this work, we introduce a domain-, model-, explanation-agnostic, generalizable and reproducible framework that ensures both transparency and human-centered explanations tailored to the needs of both experts and non-experts. The framework leverages Large Language Models (LLMs) and employs in-context learning to convey domain- and explainability-relevant contextual knowledge into LLMs. Through its structured prompt and system setting, our framework encapsulates in one response explanations understandable by non-experts and technical information to experts, all grounded in domain and explainability principles. To demonstrate the effectiveness of our framework, we establish a ground-truth contextual ``thesaurus'' through a rigorous benchmarking with over 40 data, model, and XAI combinations for an explainable clustering analysis of a well-being scenario. Through a comprehensive quality and human-friendliness evaluation of our framework's explanations, we prove high content quality through strong correlations with ground-truth explanations (Spearman rank correlation=0.92) and improved interpretability and human-friendliness to non-experts through a user study (N=56). Our overall evaluation confirms trust in LLMs as HCXAI enablers, as our framework bridges the above Gaps by delivering (i) high-quality technical explanations aligned with foundational XAI methods and (ii) clear, efficient, and interpretable human-centered explanations for non-experts.

Figures

Figures reproduced from arXiv: 2506.12240 by the authors.

Figure 1
Figure 1. The overview of our work. 3.1 Application Domain & Data As mentioned in section 1, we demonstrate the applicability of our framework in the critical healthcare domain, where democratizing XAI is essential not only for compliance with current legal standards but also to ensure that non-experts (i.e. doctors and patients) receive human-centered explanations, thereby increas￾ing engagement and trust. We focus on the co… view at source ↗
Figure 2
Figure 2. Overview of the individual’s characteristics in the two clusters -negative [PITH_FULL_IMAGE:figures/full_fig_p008_2.png] view at source ↗
Figure 3
Figure 3. The visualization of the feature￾importance-based LIME explanation of a randomly selected data point. Tab. 1 summarizes the average explanation quality for each XAI method, based on a sample of 20 randomly selected data points. For local methods (i.e. LIME, Anchor, Counterfactuals), quality is assessed at the instance level and averaged across the 20 data points, while for the global method (i.e. Coefficients), qual… view at source ↗
Figures from the paper (3 more)
Figure 4
Figure 4. Figure 4: The high-level architecture of our framework. [PITH_FULL_IMAGE:figures/full_fig_p010_4.png]
Figure 5
Figure 5. Figure 5: The generalizable and configurable templates of the prompt (5a) and [PITH_FULL_IMAGE:figures/full_fig_p011_5.png]
Figure 6
Figure 6. Figure 6: An indicative output human-centered explanation of our framework. [PITH_FULL_IMAGE:figures/full_fig_p012_6.png]

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

67 extracted references · 47 canonical work pages

  1. [1]

    Concept Induction using LLMs: a user experiment for assessment

    Barua,A.,Widmer,C.,Hitzler,P.:Conceptinductionusingllms:auserexperiment for assessment. arXiv preprint arXiv:2404.11875 (2024)

    work page Pith review arXiv 2024

  2. [2]

    Communications of the ACM67(12), 58–65 (2024)

    Bellogín, A., Grau, O., Larsson, S., Schimpf, G., Sengupta, B., Solmaz, G.: The eu ai act and the wager on trustworthy ai. Communications of the ACM67(12), 58–65 (2024)

    work page 2024

  3. [3]

    Computers & geosciences10(2-3), 191–203 (1984)

    Bezdek, J.C., Ehrlich, R., Full, W.: Fcm: The fuzzy c-means clustering algorithm. Computers & geosciences10(2-3), 191–203 (1984)

    work page 1984

  4. [4]

    Bhattacharjee, A., Moraffah, R., Garland, J., Liu, H.: Towards llm-guided causal explainability for black-box text classifiers (2024)

    work page 2024

  5. [5]

    IEEE Access10, 101556– 101574 (2022)

    Bobek, S., Kuk, M., Szelążek, M., Nalepa, G.J.: Enhancing cluster analysis with explainable ai and multidimensional cluster prototypes. IEEE Access10, 101556– 101574 (2022)

    work page 2022

  6. [6]

    Advances in neural information processing systems33, 1877–1901 (2020)

    Brown, T., Mann, B., Ryder, N., Subbiah, M., Kaplan, J.D., Dhariwal, P., Nee- lakantan, A., Shyam, P., Sastry, G., Askell, A., et al.: Language models are few-shot learners. Advances in neural information processing systems33, 1877–1901 (2020)

    2020

  7. [7]

    at-Automatisierungstechnik70(9), 787–792 (2022)

    Burkart, N., Brajovic, D., Huber, M.F.: Explainable ai: introducing trust and com- prehensibility to ai engineering. at-Automatisierungstechnik70(9), 787–792 (2022)

    work page 2022

  8. [8]

    arXiv preprint arXiv:2403.16512 (2024)

    Cahyawijaya, S., Lovenia, H., Fung, P.: Llms are few-shot in-context low-resource language learners. arXiv preprint arXiv:2403.16512 (2024)

    arXiv 2024

Show all 67 references
  1. [9]

    Communications in Statistics-theory and Methods3(1), 1–27 (1974)

    Caliński, T., Harabasz, J.: A dendrite method for cluster analysis. Communications in Statistics-theory and Methods3(1), 1–27 (1974)

    1974

  2. [10]

    In: Pacific-Asia conference on knowledge discovery and data mining

    Campello, R.J., Moulavi, D., Sander, J.: Density-based clustering based on hier- archical density estimates. In: Pacific-Asia conference on knowledge discovery and data mining. pp. 160–172. Springer (2013)

    2013

  3. [11]

    Machine Learning and Knowledge Extraction5(1), 78–108 (2023)

    Clement, T., Kemmerzell, N., Abdelaal, M., Amberg, M.: Xair: A systematic metareview of explainable ai (xai) aligned to the software development process. Machine Learning and Knowledge Extraction5(1), 78–108 (2023)

    2023

  4. [12]

    IEEE transactions on pattern analysis and machine intelligence (2), 224–227 (1979)

    Davies, D.L., Bouldin, D.W.: A cluster separation measure. IEEE transactions on pattern analysis and machine intelligence (2), 224–227 (1979)

    1979

  5. [13]

    In: Proceedings of the 2021 ACM Designing Interactive Systems Conference

    Dhanorkar, S., Wolf, C.T., Qian, K., Xu, A., Popa, L., Li, Y.: Who needs to know what, when?: Broadening the explainable ai (xai) design space by looking at explanations across the ai lifecycle. In: Proceedings of the 2021 ACM Designing Interactive Systems Conference. pp. 1591...

    2021

  6. [14]

    International Journal of Automation and Computing17(6), 788–811 (2020)

    Domański, P.D.: Study on statistical outlier detection and labelling. International Journal of Automation and Computing17(6), 788–811 (2020)

    2020

  7. [15]

    Information Sciences571, 375–400 (2021)

    Du, M., Wang, R., Ji, R., Wang, X., Dong, Y.: Robp a robust border-peeling clustering using cauchy kernel. Information Sciences571, 375–400 (2021)

    2021

  8. [16]

    Journal of cy- bernetics 4(1), 95–104 (1974)

    Dunn, J.C.: Well-separated clusters and optimal fuzzy partitions. Journal of cy- bernetics 4(1), 95–104 (1974)

    1974

  9. [17]

    Ester, M., Kriegel, H.P., Sander, J., Xu, X., et al.: A density-based algorithm for discovering clusters in large spatial databases with noise. In: kdd. vol. 96, pp. 226–231 (1996)

    1996

  10. [18]

    In: Proceedings of the 39th International 22 E

    Fang, A., Macdonald, C., Ounis, I., Habel, P.: Using word embedding to evaluate the coherence of topics from twitter data. In: Proceedings of the 39th International 22 E. Paraschou et al. ACM SIGIR conference on Research and Development in Information Retrieval. pp. 1057–1060 (2016)

    2016

  11. [19]

    Fang, X., Xu, W., Tan, F.A., Zhang, J., Hu, Z., Qi, Y.J., Nickleach, S., Socolinsky, D., Sengamedu, S., Faloutsos, C., et al.: Large language models (llms) on tabular data: Prediction, generation, and understanding-a survey (2024)

    2024

  12. [20]

    JAMA internal medicine 182(7), 730–738 (2022)

    Fawzy, A., Wu, T.D., Wang, K., Robinson, M.L., Farha, J., Bradke, A., Golden, S.H., Xu, Y., Garibaldi, B.T.: Racial and ethnic discrepancy in pulse oximetry and delayed identification of treatment eligibility among patients with covid-19. JAMA internal medicine 182(7), 730–738 (2022)

    2022

  13. [21]

    Retrieved on 30 (2016)

    Funk, M., et al.: Global burden of mental disorders and the need for a compre- hensive, coordinated response from health and social sectors at the country level. Retrieved on 30 (2016)

    2016

  14. [22]

    O’Reilly Media, Inc

    Géron, A.: Hands-on machine learning with Scikit-Learn, Keras, and TensorFlow. " O’Reilly Media, Inc." (2022)

    2022

  15. [23]

    Psy- chological assessment 4(1), 26 (1992)

    Goldberg, L.R.: The development of markers for the big-five factor structure. Psy- chological assessment 4(1), 26 (1992)

    1992

  16. [24]

    In: The Twelfth International Conference on Learning Representations (2024), https://openreview.net/forum?id¯jE8xbmvFin

    Gurnee, W., Tegmark, M.: Language models represent space and time. In: The Twelfth International Conference on Learning Representations (2024), https://openreview.net/forum?id¯jE8xbmvFin

    2024

  17. [25]

    Frontiers in public health8, 357 (2020)

    Iwendi, C., Bashir, A.K., Peshkar, A., Sujatha, R., Chatterjee, J.M., Pasupuleti, S., Mishra, R., Pillai, S., Jo, O.: Covid-19 patient health prediction using boosted random forest algorithm. Frontiers in public health8, 357 (2020)

    2020

  18. [26]

    ACM Transactions on Information Systems (TOIS)20(4), 422–446 (2002)

    Järvelin, K., Kekäläinen, J.: Cumulated gain-based evaluation of ir techniques. ACM Transactions on Information Systems (TOIS)20(4), 422–446 (2002)

    2002

  19. [27]

    In: Proceedings of the 20th In- ternational Conference on Human-Computer Interaction with Mobile De- vices and Services

    Katevas, K., Arapakis, I., Pielot, M.: Typical phone use habits: intense use does not predict negative well-being. In: Proceedings of the 20th In- ternational Conference on Human-Computer Interaction with Mobile De- vices and Services. MobileHCI ’18, Association for Computing ...

    2018

  20. [28]

    Laugwitz, B., Held, T., Schrepp, M.: Construction and evaluation of a user experi- ence questionnaire. In: HCI and Usability for Education and Work: 4th Symposium of the Workgroup Human-Computer Interaction and Usability Engineering of the Austrian Computer Society, USAB 2008,...

    2008

  21. [29]

    nature 521(7553), 436–444 (2015)

    LeCun, Y., Bengio, Y., Hinton, G.: Deep learning. nature 521(7553), 436–444 (2015)

    2015

  22. [30]

    In: The Eleventh International Conference on Learning Representations (2023), https://openreview.net/forum?id¯DeG07_TcZvT

    Li, K., Hopkins, A.K., Bau, D., Viégas, F., Pfister, H., Wattenberg, M.: Emergent world representations: Exploring a sequence model trained on a synthetic task. In: The Eleventh International Conference on Learning Representations (2023), https://openreview.net/forum?id¯DeG07_TcZvT

    2023

  23. [31]

    arXiv preprint arXiv:2110.10790 (2021)

    Liao, Q.V., Varshney, K.R.: Human-centered explainable ai (xai): From algorithms to user experiences. arXiv preprint arXiv:2110.10790 (2021)

    2021 arXiv

  24. [32]

    arXiv preprint arXiv:2305.15525 (2023)

    Liu, X., McDuff, D., Kovacs, G., Galatzer-Levy, I., Sunshine, J., Zhan, J., Poh, M.Z., Liao, S., Di Achille, P., Patel, S.: Large language models are few-shot health learners. arXiv preprint arXiv:2305.15525 (2023)

    2023 arXiv

  25. [33]

    European journal of pain 25(2), 442–465 (2021) Mind the XAI Gap 23

    Loetsch, J., Malkusch, S.: Interpretation of cluster structures in pain-related phe- notype data using explainable artificial intelligence (xai). European journal of pain 25(2), 442–465 (2021) Mind the XAI Gap 23

    2021

  26. [34]

    IEEE (2023)

    Ma, L., Thakurdesai, N., Chen, J., Xu, J., Korpeoglu, E., Kumar, S., Achan, K.: Llms with user-defined prompts as generic data operators for reliable data process- ing.In:2023IEEEInternationalConferenceonBigData(BigData).pp.3144–3148. IEEE (2023)

    2023

  27. [35]

    arXiv preprint arXiv:2406.02377 (2024)

    Ma, Q., Ren, X., Huang, C.: Xrec: Large language models for explainable recom- mendation. arXiv preprint arXiv:2406.02377 (2024)

    2024 arXiv

  28. [36]

    In: Proceedings of the fifth Berkeley symposium on mathematical statistics and probability

    MacQueen, J., et al.: Some methods for classification and analysis of multivariate observations. In: Proceedings of the fifth Berkeley symposium on mathematical statistics and probability. vol. 1, pp. 281–297. Oakland, CA, USA (1967)

    1967

  29. [37]

    Mavrepis, P., Makridis, G., Fatouros, G., Koukos, V., Separdani, M.M., Kyriazis, D.: Xai for all: Can large language models simplify explainable ai? arXiv preprint arXiv:2401.13110 (2024)

    2024 arXiv

  30. [38]

    Molnar, C.: Interpretable machine learning. Lulu. com (2020)

    2020

  31. [39]

    Tutorials in quantitative Methods for Psychology 4(1), 13–20 (2008)

    Nachar, N., et al.: The mann-whitney u: A test for assessing whether two indepen- dent samples come from the same distribution. Tutorials in quantitative Methods for Psychology 4(1), 13–20 (2008)

    2008

  32. [40]

    Advances in neural information processing sys- tems 35, 27730–27744 (2022)

    Ouyang, L., Wu, J., Jiang, X., Almeida, D., Wainwright, C., Mishkin, P., Zhang, C., Agarwal, S., Slama, K., Ray, A., et al.: Training language models to follow instructions with human feedback. Advances in neural information processing sys- tems 35, 27730–27744 (2022)

    2022

  33. [41]

    International Journal of Human– Computer Interaction 39(3), 391–437 (2023)

    Ozmen Garibay, O., Winslow, B., Andolina, S., Antona, M., Bodenschatz, A., Coursaris, C., Falco, G., Fiore, S.M., Garibay, I., Grieman, K., et al.: Six human- centered artificial intelligence grand challenges. International Journal of Human– Computer Interaction 39(3), 391–437 (2023)

    2023

  34. [42]

    Pattern recognition37(3), 487–501 (2004)

    Pakhira, M.K., Bandyopadhyay, S., Maulik, U.: Validity index for crisp and fuzzy clusters. Pattern recognition37(3), 487–501 (2004)

    2004

  35. [43]

    arXiv preprint arXiv:2402.03366 (2024)

    Peng, Y., Chen, H., Lin, C., Huang, G., Hu, J., Guo, H., Kong, B., Hu, S., Wu, X., Wang, X.: Uncertainty-aware explainable recommendation with large language models. arXiv preprint arXiv:2402.03366 (2024)

    2024 arXiv

  36. [44]

    In: Proceedings of the 48th annual meeting of the Association for Computational Linguistics

    Pitler, E., Louis, A., Nenkova, A.: Automatic evaluation of linguistic quality in multi-document summarization. In: Proceedings of the 48th annual meeting of the Association for Computational Linguistics. pp. 544–554 (2010)

    2010

  37. [45]

    arXiv preprint arXiv:2305.05138 (2023)

    Qin, W., Chen, Z., Wang, L., Lan, Y., Ren, W., Hong, R.: Read, diagnose and chat: Towards explainable and interactive llms-augmented depression detection in social media. arXiv preprint arXiv:2305.05138 (2023)

    2023 arXiv

  38. [46]

    Ramlochan, S.: The black box problem: Opaque inner workings of large language models. Prompt Engineering (2024), https://promptengineering.org/the-black- box-problem-opaque-inner-workings-of-large-language-models/#what-is-the-llm- black-box-problem, accessed: 2024-06-29

    2024

  39. [47]

    why should i trust you?

    Ribeiro, M.T., Singh, S., Guestrin, C.: " why should i trust you?" explaining the predictions of any classifier. In: Proceedings of the 22nd ACM SIGKDD interna- tional conference on knowledge discovery and data mining. pp. 1135–1144 (2016)

    2016

  40. [48]

    In: Proceedings of the AAAI conference on artificial intelligence

    Ribeiro, M.T., Singh, S., Guestrin, C.: Anchors: High-precision model-agnostic explanations. In: Proceedings of the AAAI conference on artificial intelligence. vol. 32 (2018)

    2018

  41. [49]

    Journal of computational and applied mathematics20, 53–65 (1987)

    Rousseeuw, P.J.: Silhouettes: a graphical aid to the interpretation and validation of cluster analysis. Journal of computational and applied mathematics20, 53–65 (1987)

    1987

  42. [50]

    arXiv preprint arXiv:2302.11577 (2023) 24 E

    Rozario, S., Čevora, G.: Explainable ai does not provide the explanations end-users are asking for. arXiv preprint arXiv:2302.11577 (2023) 24 E. Paraschou et al

    2023 arXiv

  43. [51]

    Bmj349 (2014)

    Sedgwick, P.: Spearman’s rank correlation coefficient. Bmj349 (2014)

    2014

  44. [52]

    Senter, R., Smith, E.A.: Automated readability index. Tech. rep., Technical report, DTIC document (1967)

    1967

  45. [53]

    Nature medicine27(12), 2176–2182 (2021)

    Seyyed-Kalantari, L., Zhang, H., McDermott, M.B., Chen, I.Y., Ghassemi, M.: Underdiagnosisbiasofartificialintelligencealgorithmsappliedtochestradiographs in under-served patient populations. Nature medicine27(12), 2176–2182 (2021)

    2021

  46. [54]

    IEEE Transactions on pattern analysis and machine intelligence22(8), 888–905 (2000)

    Shi, J., Malik, J.: Normalized cuts and image segmentation. IEEE Transactions on pattern analysis and machine intelligence22(8), 888–905 (2000)

    2000

  47. [55]

    (1999), https://api.semanticscholar.org/CorpusID:150086849

    Spielberger, C.D., Sydeman, S.J., Owen, A., Marsh, B.J.: Mea- suring anxiety and anger with the state-trait anxiety inventory (stai) and the state-trait anger expression inventory (staxi). (1999), https://api.semanticscholar.org/CorpusID:150086849

    1999

  48. [56]

    arXiv preprint arXiv:2403.07940 (2024)

    Sultanpure, K., Shirsath, B., Bhande, B., Sawai, H., Gawade, S., Samgir, S.: Hair and scalp disease detection using deep learning. arXiv preprint arXiv:2403.07940 (2024)

    2024 arXiv

  49. [57]

    arXiv preprint arXiv:2307.00364 (2023)

    Swamy, V., Frej, J., Käser, T.: The future of human-centric explainable artificial intelligence (xai) is not post-hoc explanations. arXiv preprint arXiv:2307.00364 (2023)

    2023 arXiv

  50. [58]

    In: Proceedings of the NAACL HLT 2010 student research workshop

    Vadlapudi, R., Katragadda, R.: On automated evaluation of readability of sum- maries: Capturing grammaticality, focus, structure and coherence. In: Proceedings of the NAACL HLT 2010 student research workshop. pp. 7–12 (2010)

    2010

  51. [59]

    Wachter, S., Mittelstadt, B., Russell, C.: Counterfactual explanations without opening the black box: Automated decisions and the gdpr. Harv. JL & Tech.31, 841 (2017)

    2017

  52. [60]

    Applied Physiology, Nutrition, and Metabolism36(S1), S266–S298 (2011)

    Warburton, D.E., Gledhill, N., Jamnik, V.K., Bredin, S.S., McKenzie, D.C., Stone, J., Charlesworth, S., Shephard, R.J.: Evidence-based risk assessment and recom- mendations for physical activity clearance: Consensus document 2011. Applied Physiology, Nutrition, and Metabolism3...

    2011

  53. [61]

    Journal of personality and social psychology 54(6), 1063 (1988)

    Watson, D., Clark, L.A., Tellegen, A.: Development and validation of brief mea- sures of positive and negative affect: the panas scales. Journal of personality and social psychology 54(6), 1063 (1988)

    1988

  54. [62]

    Information Fusion92, 154–176 (2023)

    Weber, L., Lapuschkin, S., Binder, A., Samek, W.: Beyond explaining: Oppor- tunities and challenges of xai-based model improvement. Information Fusion92, 154–176 (2023)

    2023

  55. [63]

    IEEE Transactions on Pattern Analysis & Machine Intelligence13(08), 841–847 (1991)

    Xie, X.L., Beni, G.: A validity measure for fuzzy clustering. IEEE Transactions on Pattern Analysis & Machine Intelligence13(08), 841–847 (1991)

    1991

  56. [64]

    ACM Transactions on Knowledge Discovery from Data18(6), 1–32 (2024)

    Yang, J., Jin, H., Tang, R., Han, X., Feng, Q., Jiang, H., Zhong, S., Yin, B., Hu, X.: Harnessing the power of llms in practice: A survey on chatgpt and beyond. ACM Transactions on Knowledge Discovery from Data18(6), 1–32 (2024)

    2024

  57. [65]

    Scientific Data9(1), 663 (2022)

    Yfantidou, S., Karagianni, C., Efstathiou, S., Vakali, A., Palotti, J., Giakatos, D.P., Marchioro, T., Kazlouski, A., Ferrari, E., Girdzijauskas, Š.: Lifesnaps, a 4- month multi-modal dataset capturing unobtrusive snapshots of our lives in the wild. Scientific Data9(1), 663 (2022)

    2022

  58. [66]

    arXiv preprint arXiv:2306.11025 (2023)

    Yu, X., Chen, Z., Ling, Y., Dong, S., Liu, Z., Lu, Y.: Temporal data meets llm–explainable financial time series forecasting. arXiv preprint arXiv:2306.11025 (2023)

    2023 arXiv

  59. [67]

    arXiv preprint arXiv:2309.10346 (2023)

    Zhang, X., Guo, Y., Stepputtis, S., Sycara, K., Campbell, J.: Explaining agent behavior with large language models. arXiv preprint arXiv:2309.10346 (2023)

    2023 arXiv

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.