Pith. sign in

REVIEW 4 major objections 6 minor 33 references

Holistix: A Dataset for Holistic Wellness Dimensions Analysis in Mental Health Narratives

T0 review · 4 major / 6 minor · reviewed 2026-08-06 · deepseek-v4-flash

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

Pith's one-line read Holistix introduces 1,420 forum posts labeled for six wellness dimensions, with the best model reaching 74% accuracy.

desk verdict A useful new wellness-dimension dataset for mental-health NLP, but the gold-label pipeline and evaluation protocol need to be spelled out before the benchmark numbers are trusted. read the letter →

arxiv 2507.09565 v2 pith:A5OZ3ESF submitted 2025-07-13 cs.LG

Heba Shakeel , Tanvir Ahmad , Chandni Saxena This is my paper · ORCID
classification cs.LG
keywords wellnessdimensionsclassificationmentalhealthnarrativessocialmediadatasetBeyondBlueforumtextspanannotationmulti-classBERTexplainability
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 introduces Holistix, a dataset of 1,420 posts from Australia's Beyond Blue mental health forums, each labeled by trained annotators into one of six wellness dimensions: physical, emotional, social, intellectual, spiritual, and vocational, together with the text spans that justify the label. The aim is to make mental-health assessment from social media more fine-grained: instead of predicting a binary crisis or not, a model can identify which dimension of a person's life is under strain. The paper reports that a fine-tuned MentalBERT model reaches 74% accuracy over ten-fold cross-validation, with strong F1 scores for vocational, social, and physical categories and much weaker ones for emotional and spiritual categories. It also evaluates LIME explanations against the annotated spans, connecting the classification to a transparent, human-readable justification. A sympathetic reading is that the dataset, released publicly, provides a region-specific resource for moving from crude mental-health signals toward dimension-level, explainable assessment.

What carries the argument

The load-bearing object is the annotation framework: a six-class label scheme derived from Dunn's theory of high-level wellness and Hettler's Six Dimensions model, operationalized through a table of class indicators (for example, fatigue and sleep for physical, career dissatisfaction for vocational, loneliness for social) and a set of perplexity guidelines for resolving overlap, ambiguity, and metaphor. Two student annotators applied this framework independently, with a Fleiss kappa of 0.7592, and the expert-developed guidelines supplied the indicators and resolved ambiguous cases. This framework does the work of turning raw forum text into a supervised learning signal; the classifier scores and the explanation-span evaluation are both measured against it.

What would settle it

Re-annotate the 223 emotional posts and a sample of the 190 spiritual posts with a fresh group of clinical psychologists using the same published guidelines; if their labels agree with the released labels only at chance levels (for example, kappa below 0.4), or if the disagreement concentrates on the same ambiguous phrases flagged in the limitations, then the reported F1 scores for those dimensions measure label noise rather than model skill.

Watch

Extended reading notes

Core claim

The paper's central claim is that wellness-dimension classification with explanatory span annotation is a viable, finer-grained task for mental-health text, and that the Holistix dataset supports it. On its own terms, the contribution is the dataset itself: 1,420 preprocessed posts from Beyond Blue forums, annotated under expert-designed guidelines grounded in Dunn and Hettler's six-dimension model, with each post assigned a dominant wellness aspect and key text spans highlighted. The evaluation establishes that transformer models fine-tuned on this dataset clearly outperform classical TF-IDF classifiers (74% versus 52% best accuracy), that MentalBERT is the strongest baseline, and that emotional and spiritual aspects are the hardest classes across all models, with MentalBERT's emotional F1 at 0.48. The paper further claims that LIME-generated explanations align with the human span annotations, with MentalBERT ahead of logistic regression on F1, ROUGE, and BLEU, showing that the span labels are not just decoration but a measurable explanation target.

Load-bearing premise

The benchmark stands or falls on whether the expert-defined wellness taxonomy correctly captures the distress expressed in these posts, and on whether the two student annotators' labels, with agreement of 0.7592, are accurate enough to be treated as ground truth for training and evaluation.

Editorial extensions

If this is right

  • The public Holistix dataset gives researchers a region-specific benchmark for classifying wellness dimensions in social-media mental-health text, with explanatory spans that go beyond simple class labels.
  • Since transformer models clearly beat classical TF-IDF classifiers, future work on this task can reasonably focus on pretrained or instruction-tuned language models rather than feature engineering.
  • The consistently low F1 scores for emotional and spiritual dimensions across all models indicate that these two classes need either more training data, different label definitions, or alternative representations before they are reliable enough for deployment.
  • The span annotations backed by LIME similarity metrics provide a concrete way to evaluate whether a classifier's justification matches human reasoning, not just whether its predicted label is correct.
  • The paper's stated next step of multi-label classification follows directly from its own perplexity guidelines, which instruct annotators to label all relevant dimensions and highlight the dominant one.

Reading between the lines

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

  • The single-label design may undercount the co-occurring dimensions that the perplexity guidelines explicitly acknowledge, so a multi-label version of the same corpus could yield higher measured performance and better reflect the reality of mental-health narratives.
  • The difficulty with emotional and spiritual classes may be an annotation-subjectivity problem as much as a model deficiency; treating those labels as soft or uncertainty-weighted could be more informative than forcing a single dominant class.
  • Because the data come from one Australian forum, models trained on Holistix may not transfer to other regions or platforms without fine-tuning, making the dataset a natural test bed for cross-cultural mental-health NLP adaptation.
  • The annotated spans could be used to train extractive explanation models directly, rather than only evaluating post-hoc LIME output, which would turn explanation generation into a supervised task on this resource.
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

4 major / 6 minor

Summary. The paper introduces Holistix, a dataset of 1,420 posts from Australia's Beyond Blue mental health forums, each annotated with one of six wellness dimensions (physical, emotional, social, intellectual, spiritual, vocational) together with explanatory text spans. The authors describe an expert-developed annotation framework and report an overall Fleiss' kappa of 75.92%. They then evaluate classical machine-learning baselines (logistic regression, linear SVM, Gaussian NB) and fine-tuned transformer models (BERT, DistilBERT, MentalBERT, Flan-T5, XLNet, GPT-2) for single-label multi-class classification, with MentalBERT achieving the highest reported accuracy of 74%. The paper also applies LIME to MentalBERT and logistic regression and reports ROUGE, BLEU, precision, recall, and F1 scores for the resulting explanations against the annotated spans. The central claims are that Holistix is a reliable, reusable wellness-dimension benchmark for the Australian population and that the reported baselines establish a meaningful performance landscape.

Significance. If the gold labels are trustworthy, Holistix would fill a concrete gap: it is a publicly released, expert-grounded dataset covering all six wellness dimensions in a region-specific population, extending prior four-dimension resources such as Wellxplain. The paper provides a clear taxonomy, worked indicator examples, and open code and data on GitHub, which are good signs for reproducibility. However, the value of the benchmark hinges on the annotation pipeline: two load-bearing gaps—the unresolved final-label adjudication rule and the absence of per-class agreement—mean that the reported per-class F1 scores cannot yet be interpreted as measurements against a reliable gold standard. The evaluation-protocol ambiguity and missing statistical uncertainty further weaken the model-comparison claims. These issues are fixable within the manuscript's scope, but they must be resolved before the dataset can serve as a dependable benchmarking resource.

major comments (4)
  1. [Section III, first paragraph; Table IV] The manuscript states: 'We used a fixed set of 990 training samples, 212 validation samples, and 213 test samples, with performance evaluated using 10-fold cross-validation.' A fixed split and 10-fold cross-validation are mutually incompatible evaluation protocols, yet Table IV is captioned as averaged over 10 folds and the abstract repeats the 10-fold claim. Please specify which protocol produced the reported numbers, how the validation set is used during hyperparameter selection, and how the 1,420 posts are partitioned across folds. Without this clarification, the headline 74% MentalBERT accuracy and the per-class F1 scores are not reproducible.
  2. [Section II-E and Section II-D.1, guideline 7] The final-label pipeline is underspecified. Section II-E says two student annotators 'annotate the data instances independently,' but guideline 7 in Section II-D.1 mentions only that quality can be checked by having a second annotator 'review 20% of the entries.' The paper never states whether both annotators labeled all 1,420 posts or whether the second annotator reviewed only a subset, and no adjudication step is described for converting the annotators' labels into the single gold label per post implied by Table II. Because Section IV itself concedes that emotional and spiritual dimensions produce divergent interpretations of the same text, an explicit rule for resolving disagreements and per-class inter-annotator agreement are needed before the Table IV per-class F1 values can be treated as measuring a known-good target.
  3. [Table IV and Section III-B] All classification results are reported as point estimates with no standard deviations, confidence intervals, or significance tests. With per-class sample sizes between 150 and 406, differences such as MentalBERT's Vocational F1 of 0.87 versus DistilBERT's 0.79, or BERT's Emotional F1 of 0.39 versus MentalBERT's 0.48, may well be within fold-to-fold noise. The paper should report per-fold variance and, ideally, paired significance tests before claiming that MentalBERT 'is the top choice' and that DistilBERT and GPT-2 offer 'strong performance' on the basis of these numbers.
  4. [Section III-B, Explainability Evaluation; Table V] The LIME-span evaluation is too underspecified to interpret. The sentence 'calculate the similarity score between the LIME-generated predictions and the annotated explanation spans using keywords' does not say how LIME's feature-level explanations are aligned with the annotated spans, how keywords are matched (exact tokens, stemming, partial overlap), or what the units of comparison are. ROUGE and BLEU are usually defined for generated text versus a reference, so the exact input format must be described. As written, the claim that MentalBERT outperforms logistic regression across all interpretability metrics is not reproducible.
minor comments (6)
  1. [Section II-D heading] The term 'Perplexity' appears to be used where 'ambiguity' is meant; please define the term explicitly or replace it.
  2. [Table III] The column header 'Avg. Count' appears to list raw frequencies (e.g., job(45)), not average counts; please clarify or relabel the column.
  3. [Section III-A] For reproducibility, report the transformer fine-tuning details: optimizer, warmup steps, maximum sequence length, number of runs, and early-stopping criterion, and explain why the BERT-family learning rate is set to 1e-3.
  4. [Section II-C] The interpretation that the lower counts of IA and VA 'may indicate that these areas are well-regulated and efficiently managed' is an unsupported causal claim; please soften it or provide supporting evidence.
  5. [Section II-A] Please specify the date range of the scraped posts, the exact selection criteria for threads and subforums, and whether the scraping complied with Beyond Blue's terms of service.
  6. [Section II-E] Since two annotators are used, please justify the choice of Fleiss' kappa over Cohen's kappa, or clarify that the reported value is equivalent to Cohen's kappa for the two-rater case.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: Holistix labels are expert annotations and classifier scores are held-out measurements, not fitted predictions.

full rationale

The paper's central contribution is a labeled benchmark, and its results are measurements against those labels rather than quantities derived from the model inputs by construction. The six wellness classes come from an externally grounded taxonomy (Dunn and Hettler, Section II-B.1), the gold labels are produced by human annotators guided by that taxonomy (Section II-E), and all reported accuracies and F1 scores are computed via 10-fold cross-validation against those human labels (Section III-B, Table IV). No equation defines a predicted quantity in terms of the labels, no fitted parameter is renamed as a prediction, and no load-bearing premise is justified solely by a self-citation. The citation to prior work (Wellxplain [10]) is used as background for extending four wellness dimensions to six, not as a proof that the current labels are correct; the citation to a LIME-application paper co-authored by C. Saxena [31] is a minor methodological reference for interpretability and does not force any result. Section IV candidly acknowledges annotator subjectivity for the Emotional and Spiritual classes, which is a data-quality limitation rather than circularity. The flagged annotation-pipeline gaps concern reliability and reporting completeness, not a reduction of the results to their inputs. Therefore no significant circularity is present.

Assumptions & free parameters 1 free parameters · 4 assumptions · 0 invented entities

The central claim is a new dataset plus benchmark; it rests on domain assumptions about the wellness taxonomy, the representativeness of a single forum, and the reliability of expert-guided annotations. No new physical or mathematical entities are postulated.

free parameters (1)
  • Transformer fine-tuning hyperparameters = learning rate 1e-3 (BERT/DistilBERT/MentalBERT/XLNet), 3e-4 (Flan-T5/GPT-2); batch sizes 4-16; 10 epochs
    Chosen by hand and not searched; the rankings in Table IV are conditional on these choices, and a 1e-3 learning rate for BERT-family models is unusually high.
assumptions (4)
  • domain assumption Dunn's/Hettler's six-dimension wellness model is a valid and sufficient taxonomy for annotating mental-health-related social media text.
    The entire annotation scheme in Section II-B builds on this model from prior literature; no validation is provided that the six dimensions cover the mental-health expressions in this corpus.
  • domain assumption Beyond Blue forum posts are representative enough of Australian-region wellness expressions to support region-specific claims.
    Section II-A scrapes only from one mental-health support forum; the paper treats this as a regional population sample without population-level representativeness evidence.
  • domain assumption Expert-provided class indicators and annotation guidelines yield reliable ground truth.
    Section II-B and II-D define indicators by two experts; kappa = 0.76 is reported but the final labels used for training are not adjudicated, and the two-annotator kappa procedure is ambiguous.
  • domain assumption Public forum data may be used and redistributed for research after anonymization.
    Section VI asserts ethical compliance; however, Beyond Blue's terms of service are not discussed, and the anonymization procedure is not described.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Holistix: A Dataset for Holistic Wellness Dimensions Analysis in Mental Health Narratives." pith.science (2026). https://pith.science/paper/A5OZ3ESF

@misc{pith2026250709565,
  author       = {Pith},
  title        = {Pith review of: Holistix: A Dataset for Holistic Wellness Dimensions Analysis in Mental Health Narratives},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/A5OZ3ESF}},
  note         = {Machine review of arXiv:2507.09565}
}
read the original abstract

We introduce a dataset for classifying wellness dimensions in social media user posts, covering six key aspects: physical, emotional, social, intellectual, spiritual, and vocational. The dataset is designed to capture these dimensions in user-generated content, with a comprehensive annotation framework developed under the guidance of domain experts. This framework allows for the classification of text spans into the appropriate wellness categories. We evaluate both traditional machine learning models and advanced transformer-based models for this multi-class classification task, with performance assessed using precision, recall, and F1-score, averaged over 10-fold cross-validation. Post-hoc explanations are applied to ensure the transparency and interpretability of model decisions. The proposed dataset contributes to region-specific wellness assessments in social media and paves the way for personalized well-being evaluations and early intervention strategies in mental health. We adhere to ethical considerations for constructing and releasing our experiments and dataset publicly on Github.

Figures

Figures reproduced from arXiv: 2507.09565 by the authors.

Figure 1
Figure 1. Overview of the problem formulation showing the identification of [PITH_FULL_IMAGE:figures/full_fig_p001_1.png] view at source ↗
Figure 2
Figure 2. Framework of data annotation process TABLE I CLASS INDICATORS IN SOCIAL MEDIA POSTS FOR THE ANNOTATION TASK, WHERE CLASS INDICATES THE WELLNESS DIMENSION Class Indicators in text Example PA Mentions of fatigue, sleep issues, body image concerns, diet struggles, illness, or medication. Phrases related to body shaming, physical deterioration, weight concerns, or health anxiety. “I feel exhausted all the time and can’t… view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

33 extracted references · 25 canonical work pages

  1. [1]

    World Mental Health Day 2023: mental health is a basic hu- man right,

    “World Mental Health Day 2023: mental health is a basic hu- man right,” https://www.emro.who.int/media/news/world-mental-health- day-2023-mental-health-is-a-basic-human-right.html, [Accessed 20-02- 2025]

    work page 2023

  2. [2]

    Depressive disorder (depression) — who.int,

    “Depressive disorder (depression) — who.int,” https://www.who.int/news-room/fact-sheets/detail/depression, [Accessed 20-02-2025]

    work page 2025

  3. [3]

    Mental illness stigma, help seeking, and public health programs,

    C. Henderson, S. Evans-Lacko, and G. Thornicroft, “Mental illness stigma, help seeking, and public health programs,” American journal of public health , vol. 103, no. 5, pp. 777–780, 2013

    work page 2013

  4. [4]

    Study: Social media use linked to decline in mental health — MIT Sloan — mitsloan.mit.edu,

    “Study: Social media use linked to decline in mental health — MIT Sloan — mitsloan.mit.edu,” https://mitsloan.mit.edu/ideas-made- to-matter/study-social-media-use-linked-to-decline-mental-health, [Ac- cessed 20-02-2025]

    work page 2025

  5. [5]

    Robust language-based mental health assessments in time and space through social media,

    S. Mangalik, J. C. Eichstaedt, S. Giorgi, J. Mun, F. Ahmed, G. Gill, A. V . Ganesan, S. Subrahmanya, N. Soni, S. A. Clouston et al., “Robust language-based mental health assessments in time and space through social media,” NPJ Digital Medicine , vol. 7, no. 1, p. 109, 2024

    work page 2024

  6. [6]

    Machine learning and natural language processing in mental health: systematic review,

    A. Le Glaz, Y . Haralambous, D.-H. Kim-Dufor, P. Lenca, R. Billot, T. C. Ryan, J. Marsh, J. Devylder, M. Walter, S. Berrouiguet et al., “Machine learning and natural language processing in mental health: systematic review,” Journal of medical Internet research , vol. 23, no. 5, p. e15708, 2021

    work page 2021

  7. [7]

    NLP as a Lens for Causal Analysis and Perception Mining to Infer Mental Health on Social Media

    M. Garg, C. Saxena, U. Naseem, and B. J. Dorr, “NLP as a lens for causal analysis and perception mining to infer mental health on social media,” arXiv preprint arXiv:2301.11004 , 2023

    work page Pith review arXiv 2023

  8. [8]

    Early detection of mental health crises through artifical-intelligence-powered social media analysis: A prospective observational study,

    M. A. Mansoor and K. H. Ansari, “Early detection of mental health crises through artifical-intelligence-powered social media analysis: A prospective observational study,” Journal of Personalized Medicine , vol. 14, no. 9, p. 958, 2024

    work page 2024

Show all 33 references
  1. [9]

    Mental health analysis in social media posts: a survey,

    M. Garg, “Mental health analysis in social media posts: a survey,” Archives of Computational Methods in Engineering , vol. 30, no. 3, pp. 1819–1842, 2023

    2023

  2. [10]

    Wellxplain: Wellness concept extraction and classification in reddit posts for mental health analysis,

    M. Garg, “Wellxplain: Wellness concept extraction and classification in reddit posts for mental health analysis,” Knowledge-Based Systems, vol. 284, p. 111228, 2024

    2024

  3. [11]

    Wellness: Its origins, theories and current applications in the united states,

    J. Star ´a and M. Charv ´at, “Wellness: Its origins, theories and current applications in the united states,” Acta Salus Vitae , vol. 1, no. 2, 2013

    2013

  4. [12]

    Welldunn: On the robustness and explainability of language models and large language models in identifying wellness dimensions,

    S. Mohammadi, E. Raff, J. Malekar, V . Palit, F. Ferraro, and M. Gaur, “Welldunn: On the robustness and explainability of language models and large language models in identifying wellness dimensions,” in Pro- ceedings of the 7th BlackboxNLP Workshop: Analyzing and Interpreting...

    2024

  5. [13]

    Instagram photos reveal predictive markers of depression,

    A. G. Reece and C. M. Danforth, “Instagram photos reveal predictive markers of depression,” EPJ Data Science , vol. 6, no. 1, p. 15, 2017

    2017

  6. [14]

    Public discourse and sentiment during the covid 19 pandemic: Using latent dirichlet allocation for topic modeling on twitter,

    J. Xue, J. Chen, C. Chen, C. Zheng, S. Li, and T. Zhu, “Public discourse and sentiment during the covid 19 pandemic: Using latent dirichlet allocation for topic modeling on twitter,” PloS one , vol. 15, no. 9, p. e0239441, 2020

    2020

  7. [15]

    Deep learning for suicide and depression identification with unsupervised label correction,

    A. Haque, V . Reddi, and T. Giallanza, “Deep learning for suicide and depression identification with unsupervised label correction,” in Artificial Neural Networks and Machine Learning–ICANN 2021: 30th International Conference on Artificial Neural Networks, Bratislava, Slovakia...

    2021

  8. [16]

    Assessing suicide risk and emotional distress in chinese social media: a text mining and machine learning study,

    Q. Cheng, T. M. Li, C.-L. Kwok, T. Zhu, and P. S. Yip, “Assessing suicide risk and emotional distress in chinese social media: a text mining and machine learning study,” Journal of medical internet research , vol. 19, no. 7, p. e243, 2017

    2017

  9. [17]

    Smhd: a large-scale resource for exploring online language usage for multiple mental health conditions,

    A. Cohan, B. Desmet, A. Yates, L. Soldaini, S. MacAvaney, and N. Go- harian, “Smhd: a large-scale resource for exploring online language usage for multiple mental health conditions,” in Proceedings of the 27th International Conference on Computational Linguistics, 2018, pp. 1485– 1497

    2018

  10. [18]

    Dreaddit: A reddit dataset for stress analysis in social media,

    E. Turcan and K. McKeown, “Dreaddit: A reddit dataset for stress analysis in social media,” EMNLP-IJCNLP 2019, p. 97, 2019

    2019

  11. [19]

    Cams: An annotated corpus for causal analysis of mental health issues in social media posts,

    M. Garg, C. Saxena, S. Saha, V . Krishnan, R. Joshi, and V . Mago, “Cams: An annotated corpus for causal analysis of mental health issues in social media posts,” in Proceedings of the Thirteenth Language Resources and Evaluation Conference , 2022, pp. 6387–6396

    2022

  12. [20]

    Cross-cultural differences in mental health expressions on social media,

    S. Rai, K. Shelat, D. R. Jain, K. Sivabalan, Y . M. Cho, M. Redkar, S. Sawant, and L. H. Guntuku, “Cross-cultural differences in mental health expressions on social media,” arXiv preprint arXiv:2402.11477 , 2024

    2024 arXiv

  13. [21]

    24/7 Support for Anxiety, Depression and Suicide Prevention. — beyondblue.org.au,

    “24/7 Support for Anxiety, Depression and Suicide Prevention. — beyondblue.org.au,” https://www.beyondblue.org.au, [Accessed 27-02- 2025]

    2025

  14. [22]

    Interrater reliability: the kappa statistic,

    M. L. McHugh, “Interrater reliability: the kappa statistic,” Biochemia medica, vol. 22, no. 3, pp. 276–282, 2012

    2012

  15. [23]

    T ¨ubingen-oslo at semeval-2018 task 2: Svms perform better than rnns in emoji prediction,

    C ¸ . C ¸¨oltekin and T. Rama, “T ¨ubingen-oslo at semeval-2018 task 2: Svms perform better than rnns in emoji prediction,” in Proceedings of the 12th international workshop on semantic evaluation , 2018, pp. 34–38

    2018

  16. [24]

    Semeval-2018 task 1: Affect in tweets,

    S. Mohammad, F. Bravo-Marquez, M. Salameh, and S. Kiritchenko, “Semeval-2018 task 1: Affect in tweets,” in Proceedings of the 12th international workshop on semantic evaluation , 2018, pp. 1–17

    2018

  17. [25]

    Bert: Pre-training of deep bidirectional transformers for language understanding,

    J. Devlin, M.-W. Chang, K. Lee, and K. Toutanova, “Bert: Pre-training of deep bidirectional transformers for language understanding,” in Pro- ceedings of the 2019 conference of the North American chapter of the association for computational linguistics: human language technolo...

    2019

  18. [26]

    Distilbert, a distilled version of bert: smaller, faster, cheaper and lighter,

    V . Sanh, L. Debut, J. Chaumond, and T. Wolf, “Distilbert, a distilled version of bert: smaller, faster, cheaper and lighter,” arXiv preprint arXiv:1910.01108, 2019

    1910 arXiv

  19. [27]

    Mentalbert: Publicly available pretrained language models for mental healthcare,

    S. Ji, T. Zhang, L. Ansari, J. Fu, P. Tiwari, and E. Cambria, “Mentalbert: Publicly available pretrained language models for mental healthcare,” in Proceedings of the Thirteenth Language Resources and Evaluation Conference, 2022, pp. 7184–7190

    2022

  20. [28]

    Scaling instruction-finetuned language models,

    H. W. Chung, L. Hou, S. Longpre, B. Zoph, Y . Tay, W. Fedus, Y . Li, X. Wang, M. Dehghani, S. Brahma et al., “Scaling instruction-finetuned language models,” Journal of Machine Learning Research , vol. 25, no. 70, pp. 1–53, 2024

    2024

  21. [29]

    Xlnet: Generalized autoregressive pretraining for language understanding,

    Z. Yang, Z. Dai, Y . Yang, J. Carbonell, R. R. Salakhutdinov, and Q. V . Le, “Xlnet: Generalized autoregressive pretraining for language understanding,” Advances in Neural Information Processing Systems , vol. 32, 2019

    2019

  22. [30]

    Language models are unsupervised multitask learners,

    A. Radford, J. Wu, R. Child, D. Luan, D. Amodei, I. Sutskever et al., “Language models are unsupervised multitask learners,” OpenAI blog , vol. 1, no. 8, p. 9, 2019

    2019

  23. [31]

    Explainable causal analysis of mental health on social media data,

    C. Saxena, M. Garg, and G. Ansari, “Explainable causal analysis of mental health on social media data,” in International conference on neural information processing . Springer, 2022, pp. 172–183

    2022

  24. [32]

    Rouge: A package for automatic evaluation of summaries,

    C.-Y . Lin, “Rouge: A package for automatic evaluation of summaries,” in Text summarization branches out , 2004, pp. 74–81

    2004

  25. [33]

    Bleu: a method for automatic evaluation of machine translation,

    K. Papineni, S. Roukos, T. Ward, and W.-J. Zhu, “Bleu: a method for automatic evaluation of machine translation,” in Proceedings of the 40th annual meeting of the Association for Computational Linguistics , 2002, pp. 311–318

    2002

Pith tools

Reviewed August 6, 2026 · model on record in the stance chip above.