Pith. sign in

REVIEW 4 major objections 5 minor 1 cited by

SpaLLM-Guard: Pairing SMS Spam Detection Using Open-source and Commercial LLMs

T0 review · 4 major / 5 minor · reviewed 2026-08-10 · deepseek-v4-flash

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

Pith's one-line read Fine-tuning an open-source LLM, Mixtral (8×7B), yields an SMS spam detector with 98.61% accuracy and false positive/negative rates below 2%, and the same model resists adversarial edits and concept drift far better than classical baselines.

desk verdict Useful LLM-vs-SMS-spam benchmark, but the headline fine-tuning accuracy may be inflated by train/test duplication and the few-shot numbers are overfit; needs a data-cleaning pass before the 98.61% claim is credible. read the letter →

arxiv 2501.04985 v1 pith:WKZ2X7LR submitted 2025-01-09 cs.CR

Muhammad Salman , Muhammad Ikram , Nardine Basta , Mohamed Ali Kaafar This is my paper · ORCID
classification cs.CR
keywords SMSspamdetectionlargelanguagemodelsfine-tuningfew-shotlearningchain-of-thoughtpromptingadversarialrobustnessconceptdriftMixtral
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 asks whether large language models can serve as reliable SMS spam detectors, comparing zero-shot, few-shot, chain-of-thought, and fine-tuned settings across commercial and open-source models. Its central claim is that fine-tuning makes open-source LLMs practical for this task: Mixtral (8×7B) reaches 98.61% accuracy on a balanced 5,000-spam/5,000-ham test set, with a false positive rate of 0.96% and a false negative rate of 1.85%, meeting the authors' 'Good' criteria. The study also finds that zero-shot detection is unreliable, few-shot results depend heavily on carefully selected examples and vary across models, and fine-tuned models are more resistant to adversarial text manipulations and to concept drift than classical machine-learning baselines. If the claims hold, a locally deployable open-source model could filter SMS spam with error rates balanced enough for real-world use, without sending message content to commercial APIs.

What carries the argument

The load-bearing mechanism is a comparative evaluation scaffold built on the Super Dataset (67,018 labeled SMS messages collected 2012–2023), split into a training set and a test set of 5,000 spam and 5,000 ham messages. Around this scaffold the paper layers four learning strategies: zero-shot prompting, few-shot prompting with exemplars tuned per model through an ablation study, chain-of-thought prompting, and QLoRA fine-tuning of selected open-source models. Performance is judged against three thresholds the authors define: Satisfactory (FPR and FNR below 5%), Good (below 3%), and Perfect (below 1%). Adversarial robustness is measured with six perceptible attacks implemented by extending TextAttack (spacing, character insertion, deletion, swap, substitution, and word substitution) plus three imperceptible attacks (invisible characters, homoglyphs, and bidirectional reorderings); concept drift is measured by training on the 2012 UCI dataset and testing on newer spam against five classical baselines.

What would settle it

Evaluate the same fine-tuned Mixtral checkpoint on a fresh test set of SMS messages collected after the training period, for example live 2024 carrier traffic, and check whether accuracy stays near 98.6% with false positive and false negative rates below 2%; a substantial drop would show the reported numbers are split-specific rather than a stable property of the fine-tuned model.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that parameter-efficient fine-tuning transforms open-source LLMs from mediocre zero-shot classifiers into deployable SMS spam detectors. Fine-tuning LLAMA-2 (70B), LLAMA-2 (13B), and Mixtral (8×7B) with QLoRA improves all three substantially over their zero- and few-shot results; Mixtral achieves 98.61% accuracy with a true positive rate of 98.15% and a true negative rate of 99.04%, corresponding to a false positive rate of 0.96% and a false negative rate of 1.85%, the only open-source model that keeps both errors below 3%. The paper further claims that fine-tuned Mixtral maintains 90.4% accuracy when trained on an older UCI dataset and tested on 2020–2023 spam, far above the 53–64% range of the SVM, random forest, LightGBM, LSTM, and CNN baselines under the same drift, and rises to 96.2% when five recent spam examples are added in a few-shot prompt. In adversarial settings, the fine-tuned Mixtral reaches 100% accuracy on several perceptible attacks and stays above 97% on imperceptible Unicode and homoglyph attacks.

Load-bearing premise

The reported accuracy and balanced error rates assume that the single random split of the Super Dataset into 5,000 spam and 5,000 ham test messages gives an unbiased estimate of how the fine-tuned models will behave on new SMS messages; the few-shot numbers additionally assume that selecting examples by iterating on test-set performance does not inflate them.

Editorial extensions

If this is right

  • A fine-tuned open-source LLM can be deployed locally as an SMS spam filter, avoiding the privacy and latency costs of sending messages to a commercial API.
  • Zero-shot LLM spam detection is not reliable enough for deployment, and few-shot performance without careful per-model example selection is inconsistent.
  • Fine-tuning improves adversarial robustness, with Mixtral reaching perfect accuracy on several perceptible attack types after adaptation.
  • LLM-based spam detectors drift much less than classical models: Mixtral trained on old data and tested on new spam keeps 90.4% accuracy, versus 53–64% for conventional baselines.
  • Combining fine-tuning with few-shot examples drawn from recent spam can further close the concept-drift gap, lifting Mixtral from 90.4% to 96.2% accuracy.

Reading between the lines

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

  • If the fine-tuning numbers are stable, network operators could refresh a small set of few-shot exemplars periodically instead of full retraining; the concept-drift experiment hints at this direction but does not test a repeated update schedule.
  • The few-shot results were obtained by selecting examples after iterating on the same test set until accuracy stopped improving, so those numbers likely overstate what a practitioner would achieve without test-set access; a clean validation-split design would yield a more conservative estimate.
  • The robustness tests transform only the spam messages themselves, so they do not cover an adversary who crafts messages specifically against the fine-tuned detector; a dedicated adaptive attack could target Mixtral's blind spots.
  • A testable extension is whether smaller open-source models (7B–13B) can reach Mixtral-level balance with more data or longer fine-tuning, which would lower the hardware cost of deployment.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

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

Referee Report

4 major / 5 minor

Summary. The paper evaluates the effectiveness of several LLMs—GPT-4, DeepSeek, LLAMA-2, Mixtral, and Mistral—for SMS spam detection under zero-shot, few-shot, chain-of-thought, and fine-tuning settings. Using a combined 67,018-message dataset, the authors report that fine-tuned Mixtral (8×7B) achieves 98.61% accuracy with FPR 0.96% and FNR 1.85%, satisfying their self-defined 'Good' criteria. They also examine adversarial robustness against perceptible and imperceptible perturbations, and profile the best model under concept drift, concluding that fine-tuned LLMs are more robust than traditional ML baselines.

Significance. If the central claims hold, the paper would provide a practical demonstration that locally deployable open-source LLMs can serve as strong SMS spam detectors, with balanced error rates and improved resilience to adversarial manipulation and concept drift compared to conventional models. The paper's strengths include a broad model comparison, inclusion of both perceptible and imperceptible attack families, a dedicated concept-drift experiment, and its grounding in a recent dataset. However, the evaluation has several methodology gaps that directly affect the reliability of the reported numbers, particularly the use of the test set for few-shot example selection, the absence of de-duplication or campaign-level grouping in the train/test split, and missing fine-tuning hyperparameters and uncertainty estimates. These issues currently limit the credibility of the central accuracy claims.

major comments (4)
  1. [Section 3.3.2] The few-shot examples were selected by iterating on the evaluation set itself: "The experiment was repeated multiple times for each LLM with different SMS messages containing similar keywords until the highest accuracy was no longer increased." Consequently, the reported few-shot accuracies (Tables 3, 4, 8) are optimized on the test set and are optimistic estimates of true few-shot performance. The authors should select examples using a separate development set or via cross-validation, and then report performance on the held-out test set without further iteration.
  2. [Section 2.1 / Table 5] The central fine-tuning result rests on a single random split of the Super Dataset into 5,000 spam and 5,000 ham test messages. The paper reports no de-duplication, near-duplicate removal, or grouping by spam campaign, source, or phone number/URL. Because the Super Dataset aggregates multiple sources and real-world SMS spam is heavily templated, a message-level random split can place near-identical content in both training and test sets, potentially inflating the fine-tuned accuracy of 98.61% (Table 5). The authors should quantify the extent of train/test leakage (e.g., exact or near-duplicate matching) and report results on a leakage-free split.
  3. [Section 4 / Table 5] The fine-tuning procedure is not reproducible as reported. The paper states that QLoRA was used but does not disclose the LoRA rank, learning rate, number of epochs, batch size, quantization configuration, or any regularization/seed settings. Without these details and without confidence intervals or multiple runs, the reader cannot assess the stability of the 98.61% accuracy figure or the claim that it generalizes. The authors should provide the hyperparameters and either run multiple seeds or report bootstrap confidence intervals.
  4. [Section 5 / Table 10] The adversarial robustness evaluation in Table 10 reports accuracy only on the 300-message holdout spam set; it does not measure false positive rates on legitimate messages after perturbation. Thus the claim that LLMs are 'robust' to adversarial manipulations is only about the preservation of spam detection, not about the overall classifier behavior (e.g., whether attacks also cause ham messages to be misclassified). The confidence in this robustness claim would be much higher if the authors reported full confusion-matrix metrics on a balanced ham/spam evaluation set under each attack.
minor comments (5)
  1. [Throughout] The model name is inconsistent: 'Mistral (7B)' is sometimes written as 'Mixtral (7B)' (e.g., Table 3 rows); please correct to avoid confusion between the Mixtral mixture-of-experts model and the Mistral 7B model.
  2. [Section 2.2.3] Typo: 'finetunning' should be 'fine-tuning'.
  3. [Section 1] Typo: 'poortransferability' should be 'poor transferability'.
  4. [Section 4] The phrase 'refer to Section 5' for the train/test split should refer to Section 2.1, where the split is actually described.
  5. [Section 3.4] The text states that GPT-4 achieved a 'precision of 97.18%' in the medium-length few-shot setting; this should say 'accuracy' (as in Table 8).

Circularity Check

2 steps flagged · score 6.0 of 10

Few-shot and concept-drift scores are tuned on the evaluation data; the central fine-tuning result is self-contained.

  1. fitted input called prediction [Section 3.3.2, 'Sensitivity of LLMs to Spam Exemplars in Few-Shot Learning' (Table 3)]
    "The LLMs were then evaluated on a smaller set of 400 SMS messages (200 spam, 200 ham). The examples for each LLM were selected based on the misclassification analysis, identifying the specific types of SMS messages each LLM struggled with. The experiment was repeated multiple times for each LLM with different SMS messages containing similar keywords until the highest accuracy was no longer increased."

    The few-shot prompt is the input whose quality is being measured, but it is tuned against the same 400-message evaluation subset on which accuracy is then reported. Repeating the experiment with different exemplars 'until the highest accuracy was no longer increased' makes the Table 3 numbers the maximum of a search over prompts on the test data. These few-shot accuracies are therefore fitted to the evaluation set, not independent predictions of few-shot generalization. The fine-tuning result in Table 5 is not directly affected, but the paper's few-shot comparisons and the model-selection narrative inherit the inflated values.

  2. fitted input called prediction [Section 6, Concept Drift (Table 7)]
    "After fine-tuning, we tested the model on a new set of SMS spam messages reported between 2020 and 2023. [...] we applied few-shot learning to the fine-tuned model by selecting five representative spam examples from the most recent spam data. The results, presented in Table 7, demonstrate that this strategy significantly enhanced the model's capacity to manage concept drift."

    The concept-drift test set is the recent SMS spam corpus, and the few-shot prompt is then populated with 'five representative spam examples from the most recent spam data' — the same recent corpus on which the model is evaluated, with no separate held-out pool described. The reported improvement from 90.4% (zero-shot) to 96.2% (few-shot) is thus obtained by placing test-set content into the prompt before scoring the model on that set. The mitigation claim is forced by construction and does not demonstrate adaptation to unseen drifted spam.

full rationale

The paper's central claim — fine-tuned Mixtral reaches 98.61% accuracy with FPR 0.96% and FNR 1.85% on a random 5,000-spam/5,000-ham test split — is not circular: the model is trained on the disjoint train split and evaluated on the test split, and no parameter of that fine-tuned result is fit to the test labels. The Super Dataset is taken from the authors' prior work [35], a genuine self-citation, but it is an empirical collection and the result is an externally checkable evaluation on that data, so it does not by itself make the derivation circular. The circular steps are confined to the few-shot evaluations: Section 3.3.2 tunes exemplars on a 400-message evaluation subset until accuracy stops improving and reports those numbers as few-shot performance, and Section 6 builds the concept-drift few-shot prompt from the same recent corpus used as the test set. Both are instances of fitted inputs being reported as predictions. The lack of de-duplication in the Super Dataset is a correctness/generalization risk, not a circularity. Score 6 reflects partial circularity in secondary few-shot and concept-drift predictions, while the central fine-tuning claim remains self-contained.

Assumptions & free parameters 3 free parameters · 4 assumptions · 0 invented entities

No new particles, forces, dimensions, or theoretical entities are introduced. The paper is an empirical evaluation of existing models and datasets. The free parameters are experimental choices and undisclosed training configurations that materially affect the reported results.

free parameters (3)
  • Fine-tuning hyperparameters (QLoRA rank, learning rate, epochs, batch size, quantization) = Not disclosed
    The fine-tuning performance of Mixtral and Llama 2 depends on these choices, but the paper does not report them, so the central 98.61% accuracy cannot be reproduced or assessed for overfitting.
  • Few-shot example selection (number, category, length) = Varies by model; selected from test set
    The ablation in Section 3.3.2 chose examples by iteratively testing on the evaluation set until accuracy stopped improving. This makes the few-shot results a fitted quantity rather than an unbiased estimate.
  • Performance category thresholds (Satisfactory/Good/Perfect) = FPR/FNR < 5%, < 3%, < 1%
    These thresholds are introduced by the authors in Section 2.5 and define whether the paper declares a detector 'Satisfactory' or 'Good'. They are arbitrary and influence the headline claim that Mixtral meets 'Good' criteria.
assumptions (4)
  • domain assumption Super Dataset ground-truth labels are accurate and the random split is representative.
    The entire evaluation relies on the correctness of labels in the Super Dataset [35] and on the random 5,000/5,000 test split described in Section 2.1.
  • domain assumption API models (GPT-4, DeepSeek) produce deterministic or effectively stable outputs for the same prompts.
    The paper reports precise accuracies for commercial APIs without discussing sampling temperature or repeated calls, so the numbers are treated as if deterministic.
  • domain assumption TextAttack and the Boucher et al. attacks are valid implementations of real-world adversarial SMS perturbations.
    Section 2.4 adopts these tools as the adversarial threat model, without evaluating whether the resulting perturbed messages are realistic SMS spam.
  • domain assumption The set of 300 spam messages used for adversarial evaluation is representative of the test distribution.
    Section 2.1 creates a holdout of 300 spam messages; the robustness conclusions in Table 10 are based on this small sample, but no confidence intervals are provided.

how reviews work

0 comments
Cite this review

Pith. "Pith review of SpaLLM-Guard: Pairing SMS Spam Detection Using Open-source and Commercial LLMs." pith.science (2026). https://pith.science/paper/WKZ2X7LR

@misc{pith2026250104985,
  author       = {Pith},
  title        = {Pith review of: SpaLLM-Guard: Pairing SMS Spam Detection Using Open-source and Commercial LLMs},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/WKZ2X7LR}},
  note         = {Machine review of arXiv:2501.04985}
}
read the original abstract

The increasing threat of SMS spam, driven by evolving adversarial techniques and concept drift, calls for more robust and adaptive detection methods. In this paper, we evaluate the potential of large language models (LLMs), both open-source and commercial, for SMS spam detection, comparing their performance across zero-shot, few-shot, fine-tuning, and chain-of-thought prompting approaches. Using a comprehensive dataset of SMS messages, we assess the spam detection capabilities of prominent LLMs such as GPT-4, DeepSeek, LLAMA-2, and Mixtral. Our findings reveal that while zero-shot learning provides convenience, it is unreliable for effective spam detection. Few-shot learning, particularly with carefully selected examples, improves detection but exhibits variability across models. Fine-tuning emerges as the most effective strategy, with Mixtral achieving 98.6% accuracy and a balanced false positive and false negative rate below 2%, meeting the criteria for robust spam detection. Furthermore, we explore the resilience of these models to adversarial attacks, finding that fine-tuning significantly enhances robustness against both perceptible and imperceptible manipulations. Lastly, we investigate the impact of concept drift and demonstrate that fine-tuned LLMs, especially when combined with few-shot learning, can mitigate its effects, maintaining high performance even on evolving spam datasets. This study highlights the importance of fine-tuning and tailored learning strategies to deploy LLMs effectively for real-world SMS spam detection

Figures

Figures reproduced from arXiv: 2501.04985 by the authors.

Figure 1
Figure 1. Comparison of the highest and lowest accuracy of [PITH_FULL_IMAGE:figures/full_fig_p009_1.png] view at source ↗
Figure 2
Figure 2. Evaluation framework of LLMs using zero-shot, few-shot, fine-tuning, and chain-of-thought prompting approaches. [PITH_FULL_IMAGE:figures/full_fig_p015_2.png] view at source ↗
Figure 3
Figure 3. Accuracy of LLMs across various learning approaches. The highest accuracy for each LLM is indicated by an arrow [PITH_FULL_IMAGE:figures/full_fig_p016_3.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. POSTER: A Multi-Signal Model for Detecting Evasive Smishing

    cs.LG 2025-05 conditional novelty 4.0 of 10

    A four-stream fusion model for SMS smishing detection reports 97.89% accuracy and 0.963 F1 on 84,000 messages, but its smishing labels were created partly from a keyword list.

Reference graph

Works this paper leans on

41 extracted references · 29 canonical work pages · cited by 1 Pith paper

  1. [1]

    Action Fraud

    2023. Action Fraud. https://www.actionfraud.police.uk/. Last accessed 06 Oct 2023

    work page 2023

  2. [2]

    ACCS. 2022. ACCS Scam statistics. https://www.scamwatch.gov.au/scam- statistics. Last accessed 15 Jul 2024]

    work page 2022

  3. [3]

    Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Floren- cia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. 2023. Gpt-4 technical report. arXiv preprint arXiv:2303.08774 (2023)

    arXiv 2023

  4. [4]

    Tiago Almeida et al. 2013. Towards sms spam filtering: Results under a new dataset. JiSS 2, 1 (2013)

    work page 2013

  5. [5]

    Tiago A Almeida, José María G Hidalgo, and Akebo Yamakami. 2011. Contribu- tions to the study of SMS spam filtering: new collection and results. InProceedings of the 11th ACM symposium on Document engineering . 259–262

    2011

  6. [6]

    Nicholas Boucher, Ilia Shumailov, Ross Anderson, and Nicolas Papernot. 2022. Bad characters: Imperceptible nlp attacks. In 2022 IEEE Symposium on Security and Privacy (SP). IEEE, 1987–2004

    2022

  7. [7]

    Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. 2020. Language models are few-shot learners. Advances in neural information processing systems 33 (2020), 1877–1901

    2020

  8. [8]

    Tao Chen and Min-Yen Kan. 2013. Creating a live, public short message service corpus: the NUS SMS corpus. LRE (2013)

    work page 2013

Show all 41 references
  1. [9]

    DeepSeek. 2023. DeepSeek API. https://platform.deepseek.com/api-docs/api/ deepseek-api/. Last accessed 15 Jul 2024]

    2023

  2. [10]

    Tim Dettmers, Artidoro Pagnoni, Ari Holtzman, and Luke Zettlemoyer. 2024. Qlora: Efficient finetuning of quantized llms. Advances in Neural Information Processing Systems 36 (2024)

    2024

  3. [11]

    Trevor Dunlap, John Speed Meyers, Bradley Reaves, and William Enck. 2024. Pair- ing Security Advisories with Vulnerable Functions Using Open-Source LLMs. In International Conference on Detection of Intrusions and Malware, and Vulnerability Assessment. Springer, 350–369

    2024

  4. [12]

    Sorouralsadat Fatemi and Yuheng Hu. 2023. A Comparative Analysis of Fine- Tuned LLMs and Few-Shot Learning of LLMs for Financial Sentiment Analysis. arXiv preprint arXiv:2312.08725 (2023)

    2023 arXiv

  5. [13]

    John Fields, Kevin Chovanec, and Praveen Madiraju. 2024. A survey of text classification with transformers: How wide? how large? how long? how accurate? how expensive? how safe? IEEE Access (2024)

    2024

  6. [14]

    Yingqiang Ge, Wenyue Hua, Kai Mei, Juntao Tan, Shuyuan Xu, Zelong Li, Yongfeng Zhang, et al. 2024. Openagi: When llm meets domain experts.Advances in Neural Information Processing Systems 36 (2024)

    2024

  7. [15]

    Abdallah Ghourabi, Mahmood A Mahmood, and Qusay M Alzubi. 2020. A hybrid CNN-LSTM model for SMS spam detection in arabic and english messages.Future Internet 12, 9 (2020), 156

    2020

  8. [16]

    Florian Heimerl et al. 2014. Word cloud explorer: Text analytics based on word clouds. In ICSS

    2014

  9. [17]

    Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. 2021. Lora: Low-rank adaptation of large language models. arXiv preprint arXiv:2106.09685 (2021)

    2021 arXiv

  10. [18]

    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 . 2024. Mixtral of experts. arXiv preprint arXiv:2401.04088 (2024)

    2024 arXiv

  11. [19]

    Katikapalli Subramanyam Kalyan. 2023. A survey of GPT-3 family large language models including ChatGPT and GPT-4. Natural Language Processing Journal (2023), 100048

    2023

  12. [20]

    Yekai Li, Rufan Zhang, Wenxin Rong, and Xianghang Mi. 2024. SpamDam: To- wards Privacy-Preserving and Adversary-Resistant SMS Spam Detection. arXiv preprint arXiv:2404.09481 (2024)

    2024 arXiv

  13. [21]

    Haokun Liu, Derek Tam, Mohammed Muqeeth, Jay Mohta, Tenghao Huang, Mohit Bansal, and Colin A Raffel. 2022. Few-shot parameter-efficient fine-tuning is better and cheaper than in-context learning. Advances in Neural Information Processing Systems 35 (2022), 1950–1965

    2022

  14. [22]

    Xiaoxu Liu, Haoye Lu, and Amiya Nayak. 2021. A spam transformer model for SMS spam detection. IEEE Access 9 (2021), 80253–80263

    2021

  15. [23]

    Meta Llama. 2023. Llama 2 13B. https://huggingface.co/meta-llama/Llama-2- 13b-chat-hf. Last accessed 15 Jul 2024]

    2023

  16. [24]

    Meta Llama. 2023. Llama 2 70B. https://huggingface.co/meta-llama/Llama-2- 70b-chat-hf. Last accessed 15 Jul 2024]

    2023

  17. [25]

    Mistral AI. 2023. Model Card for Mistral7B. https://huggingface.co/mistralai/ Mistral-7B-Instruct-v0.3. Last accessed 15 Jul 2024]

    2023

  18. [26]

    Mistral AI. 2023. Model Card for Mixtral-8x7B. https://huggingface.co/mistralai/ Mixtral-8x7B-Instruct-v0.1. Last accessed 15 Jul 2024]

    2023

  19. [27]

    John X Morris, Eli Lifland, Jin Yong Yoo, Jake Grigsby, Di Jin, and Yanjun Qi

  20. [28]

    nickboucher. 2022. Imperceptible Perturbations. https://github.com/nickboucher/ imperceptible. Last accessed 15 Jul 2024]

    2022

  21. [29]

    OpenAI. 2023. Models - OpenAI API. https://platform.openai.com/docs/models/ gpt-4-turbo-and-gpt-4. Last accessed 15 Jul 2024]

    2023

  22. [30]

    OpenAI. 2024. Best practices for prompt engineering with the OpenAI API. https://help.openai.com/en/articles/6654000-best-practices-for-prompt- engineering-with-the-openai-api. Last accessed 16 Jul 2024]

    2024

  23. [31]

    OpenAI. 2024. OpenAI and Apple Announce Partnership. https://openai.com/ index/openai-and-apple-announce-partnership/. Last accessed 15 Jul 2024]

    2024

  24. [32]

    C Oswald, Sona Elza Simon, and Arnab Bhattacharya. 2022. Spotspam: Intention analysis–driven sms spam detection using bert embeddings. ACM Transactions on the Web (TWEB) 16, 3 (2022), 1–27. Salman et al

    2022

  25. [33]

    Robokiller. 2023. 2023 Mid-Year Phone Scam Report. https://www.robokiller. com/robokiller-2023-mid-year-phone-scam-report. Last accessed 15 Jul 2024]

    2023

  26. [34]

    Thaer Sahmoud and Dr Mohammad Mikki. 2022. Spam detection using BERT. arXiv preprint arXiv:2206.02443 (2022)

    2022 arXiv

  27. [35]

    Muhammad Salman, Muhammad Ikram, and Mohamed Ali Kaafar. 2024. Investi- gating Evasive Techniques in SMS Spam Filtering: A Comparative Analysis of Machine Learning Models. IEEE Access (2024)

    2024

  28. [36]

    Nilam Nur Amir Sjarif, Nurulhuda Firdaus Mohd Azmi, Suriayati Chuprat, Haslina Md Sarkan, Yazriwati Yahya, and Suriani Mohd Sam. 2019. SMS spam message detection using term frequency-inverse document frequency and ran- dom forest algorithm. Procedia Computer Science 161 (2019)...

    2019

  29. [37]

    Siyuan Tang, Xianghang Mi, Ying Li, XiaoFeng Wang, and Kai Chen. 2022. Clues in tweets: Twitter-guided discovery and analysis of SMS spam. In ACM CCS

    2022

  30. [38]

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

    2023 arXiv

  31. [39]

    Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, et al. 2022. Chain-of-thought prompting elicits reason- ing in large language models. Advances in neural information processing systems 35 (2022), 24824–24837

    2022

  32. [40]

    l” with “1

    Jingfeng Yang, Hongye Jin, Ruixiang Tang, Xiaotian Han, Qizhang Feng, Haoming Jiang, Shaochen Zhong, Bing Yin, and Xia Hu. 2024. Harnessing the power of llms in practice: A survey on chatgpt and beyond. ACM Transactions on Knowledge Discovery from Data 18, 6 (2024), 1–32. APPE...

    2024

  33. [2020]

    arXiv preprint arXiv:2005.05909 (2020)

    Textattack: A framework for adversarial attacks, data augmentation, and adversarial training in nlp. arXiv preprint arXiv:2005.05909 (2020)

    2020 arXiv

Pith tools

Reviewed August 10, 2026 · model on record in the stance chip above.