CAF-I: A Collaborative Multi-Agent Framework for Enhanced Irony Detection with Large Language Models

Mingxuan.Hu; Ziqi.Liu; Ziyang.Zhou

REVIEW 5 major objections 5 minor 33 references

CAF-I: A Collaborative Multi-Agent Framework for Enhanced Irony Detection with Large Language Models

T0 review · 5 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 Four cooperating LLM agents lift zero-shot irony detection to a new state of the art.

desk verdict Useful multi-agent framework with a clean story and ablations, but the SOTA claim trips over an unexplained 183-sample SemEval-2018 test set. read the letter →

arxiv 2506.08430 v2 pith:SUJ42WC2 submitted 2025-06-10 cs.CL cs.MA

Ziqi.Liu , Ziyang.Zhou , Mingxuan.Hu This is my paper

classification cs.CLcs.MA

keywords ironydetectionmulti-agentsystemslargelanguagemodelszero-shotclassificationchain-of-thoughtpromptingsarcasminterpretability

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 aims to show that irony detection improves when an LLM is replaced by a small team of cooperating LLM agents, each responsible for one analytical dimension. Its proposed framework, CAF-I, lets a Context Agent, a Semantic Agent, and a Rhetoric Agent each produce a judgment, share those judgments for a second round of analysis, and then hands the refined outputs to a Decision Agent, with a Refinement Evaluator Agent allowed one round of corrective feedback. On four irony and sarcasm benchmarks, the paper reports an average Macro-F1 of 76.31 percent and an average accuracy of 76.89 percent in zero-shot settings, a 4.98-point Macro-F1 improvement over the strongest prior baseline. The intended payoff is that structured collaboration compensates for the single-perspective limitations of vanilla LLM prompting and yields reasoning traces that make the classification interpretable.

What carries the argument

The load-bearing object is the agent workflow rather than any single formula. The Context Agent extracts internal context (entities, relations, and theme), optionally retrieves external knowledge, and classifies with an LLM; the Semantic Agent performs chain-of-thought analysis that separates literal meaning from implied intent, checks emotional consistency and commonsense expectations, then classifies; the Rhetoric Agent detects rhetorical devices, explains each device's function, and classifies. After an independent first pass, the three agents share their reasoning and produce second-round judgments. The Decision Agent then applies a hierarchical rule: adopt unanimous agreement, fall back to majority vote, or, on full disagreement, pick the most compelling argument. Finally, the Refinement Evaluator Agent estimates confidence and checks for contradictions, triggering at most one feedback loop that sends targeted hints back to the analysis agents. This conditional, at-most-once refinement is what the paper credits for error correction at a bounded inference cost.

What would settle it

Run CAF-I and all cited baselines on the exact same explicit test splits, using the full standard SemEval-2018 Task 3 test set instead of a possibly smaller subset, and recompute average Macro-F1; if the 4.98-point improvement shrinks or reverses when the test sets match, the central claim would not survive that check.

Watch

Extended reading notes

Core claim

On the paper's own terms, the central claim is that structured multi-agent collaboration—not a larger model or more training data—is what pushes irony detection forward. CAF-I reports the best average accuracy (76.89) and Macro-F1 (76.31) across IAC-V1, IAC-V2, MuSTARD, and SemEval-2018, surpassing fine-tuned BERT and RoBERTa as well as previous LLM prompting strategies that use chain-of-thought, contradiction, cue graphs, or cue bagging. Ablations that remove any single analysis agent or disable the refinement evaluator lower Macro-F1, which the authors interpret as evidence that each role carries weight. A robustness check with two smaller LLM backbones still shows large gains over the same backbones' basic prompting, leading the paper to claim the architecture, not the backbone, is responsible. The paper also presents a case study in which the refinement loop corrects an initial misclassification, and a quantitative test in which adding the agents' reasoning traces to a GPT-3.5 Turbo prompt raises its Macro-F1 from 68.4 to 70.2.

Load-bearing premise

The state-of-the-art claim rests on the assumption that every baseline was evaluated on the same test instances under the same settings; the paper's Table 1 lists the SemEval-2018 set as 183 samples, whereas the standard SemEval-2018 Task 3 test set is widely reported to be about 784 tweets, so the comparison may not be aligned.

Editorial extensions

If this is right

If the reported results hold, zero-shot multi-agent prompting outperforms fine-tuned BERT and RoBERTa by roughly six Macro-F1 points on average, suggesting specialized-agent collaboration can replace fine-tuning for this task.
The hierarchical decision rule and the positive ablation of every agent imply that irony detection benefits from at least three independent diagnostic perspectives (context, semantics, rhetoric), not just a single richer prompt.
Because refinement runs at most once and only when confidence is low or contradictions appear, the accuracy gain is claimed at a latency cost close to ordinary chain-of-thought prompting—about 0.8 seconds per sample more in the paper's gpt-4o-mini measurement.
The interpretability experiments imply that the agents' intermediate reasoning traces carry signal on their own: feeding them to a weaker LLM improves that model's Macro-F1 by 1.8 points.

Reading between the lines

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

A direct test of what drives the gain would be to give a single LLM the same three-perspective prompts in one call—context, semantics, rhetoric—without role separation or debate; if that equals CAF-I's accuracy, the collaboration is not the active ingredient.
Logging which examples trigger the refinement loop and comparing their before/after errors would quantify how much of the Macro-F1 gain comes from the corrective feedback rather than from the initial multi-agent agreement.
The framework's interpretability claim could be tested by showing the decision agent's synthesized justification to human raters and asking whether it makes the final label more trustworthy than a plain LLM explanation.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

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

Desk Editor's Note

read the letter

Quick take: CAF-I is a plausible and fairly well-engineered multi-agent recipe for zero-shot irony detection, and the ablations suggest each piece earns its keep. But the headline SOTA claim rests on a SemEval-2018 test set of 183 samples, which contradicts the standard 784-tweet test set, and the paper doesn't show that every baseline was run on this same subset. Until that's resolved, the 4.98-point gain is not a solid result.

What's actually new: the division into Context/Semantics/Rhetoric agents with a decision agent and one conditional refinement loop is a sensible decomposition, and as far as I can tell the specific combination is not in the cited literature. The paper is also honest about the free parameter in the refinement trigger (low confidence or contradiction), and the ablation results are messy enough to be believable—removing any agent hurts, the REAgent helps most on SemEval. The robustness check with Qwen2-7B and Llama3-8B is a good addition; the numbers show the framework transfers beyond GPT-4o.

Where it's soft: the SemEval-2018 size problem is the big one. Table 1 says 183 tweets; the standard test set for Task 3 is 784. The paper never explains where 183 comes from or verifies baseline alignment. Since CAF-I's 79.99 Ma-F1 on that set feeds directly into the 76.31 average, this is load-bearing, not a cosmetic issue. There are smaller problems: no error bars, no significance tests, no code or prompts released, and only one run at temperature 0. The explainability experiment is weak—it uses GPT-3.5 as a consumer of the stage-1 explanations, which tells you the explanations carry signal but not much about interpretability for humans. The 'aligned evaluation settings' sentence is doing a lot of work.

Net: if the SemEval-2018 subset is verified and all baselines are re-run on it, I expect the method to still beat the GPT-4o baselines on the other three datasets, but the average SOTA number would need to be recomputed. This deserves a serious referee, but the authors must supply the split details and ideally release the prompts before the claim is acceptable.

Recommendation: send it for full peer review with the dataset-alignment issue flagged as a mandatory revision. If the authors cannot explain the 183, it's a desk reject for the SOTA claim, though the framework alone is still a minor-contribution paper.

Referee Report

5 major / 5 minor

Summary. This paper proposes CAF-I, a zero-shot multi-agent LLM framework for irony detection. Three specialized agents (Context, Semantic, Rhetoric) produce first-round analyses, exchange views, and produce second-round judgments. A Decision Agent applies consensus/majority/argument-based aggregation, and a Refinement Evaluator Agent conditionally triggers at most one feedback loop. Experiments on IAC-V1, IAC-V2, MuSTARD, and SemEval-2018 compare CAF-I against LLM prompting, fine-tuned PLM, and deep learning baselines, reporting an average accuracy of 76.89% and Macro-F1 of 76.31%, a 4.98-point absolute improvement over the strongest prior baseline. Additional experiments address ablations, backbone robustness with Qwen2-7B and Llama3-8B, interpretability with GPT-3.5 Turbo, and inference efficiency.

Significance. If the empirical claims hold, the paper would provide a practical, interpretable architecture for zero-shot irony detection and a useful demonstration that structured multi-agent collaboration can outperform prompt-only LLM baselines and fine-tuned discriminative models. The design is transparent about the hand-specified decision and refinement rules, which are not fitted to test data, and the main evaluation uses external benchmark labels, so the circularity concern flagged in the stress-test note does not land. The temperature-0 setting and clear workflow are reproducibility strengths. However, because the headline SOTA claim depends on a SemEval-2018 test subset whose size is unexplained (183 vs. the canonical 784), and because no variance or significance testing is reported, the exact SOTA claim is not yet fully verified.

major comments (5)

[Section 4.1, Table 1; Section 4.2, Table 2] Table 1 reports a SemEval-2018 test size of 183 samples, whereas the canonical SemEval-2018 Task 3 test set contains 784 tweets. The paper never explains this discrepancy or states which subset was used for CAF-I and for each baseline. Since CAF-I's 79.99 Macro-F1 on SemEval-2018 contributes directly to the average 76.31 and to the claimed 4.98-point gain over GPT-4o+CoC, the headline SOTA is only meaningful if all baselines in Table 2 were measured on the same 183-item subset. The assertion of 'aligned evaluation settings' in Section 4.1 is insufficient without split-level verification; Table 3's Qwen2-7b result of 81.25 on SemEval-2018 makes this issue more salient. Please report the exact split construction and per-baseline evaluation settings, or re-run baselines on the same subset/full set.
[Section 4.2, Table 2; Abstract] The abstract and Section 4.2 describe CAF-I as achieving SOTA on the vast majority of metrics and consistently outperforming other methods, but Table 2 shows that on IAC-V2, DC-Net achieves higher Accuracy (78.00 vs. 77.87) and Macro-F1 (77.90 vs. 76.82) than CAF-I. The average-based SOTA claim may still hold, but the 'vast majority' characterization is not consistent with the paper's own table. Please qualify the claim or provide a corrected per-dataset statement.
[Tables 2, 3, and 4] No confidence intervals, standard deviations, or paired significance tests are reported; every cell appears to be a single run. Differences as small as 0.13 accuracy points on IAC-V2 are used to discuss method ordering, and the 4.98-point average Macro-F1 gap is the central quantitative claim. Without variance estimates, bootstrap CIs, or paired tests, the average-based conclusions are not statistically grounded. Please report multiple runs or appropriate significance tests for the central comparisons.
[Section 3.3, Eq. (2)] The Context Agent's workflow depends on an external search API and query formulation q = gLLM(keywords(xi)), but the paper never identifies the API, retrieval parameterization, or how often retrieval is triggered, and no ablation isolates the external-knowledge component. This makes part of the architecture unreproducible and leaves open whether retrieval, rather than multi-agent collaboration, drives a portion of the observed gains. Please specify the retrieval setup and add an ablation with retrieval disabled.
[Section 4.5] The quantitative explainability experiment lacks essential controls: it does not state which subset of SemEval-2018 is used, whether the explanations are generated by the same GPT-4o backbone as CAF-I, whether the feature prompt excludes the final label and judgment (to avoid leakage), or whether the Macro-F1 increase from 68.4% to 70.2% is statistically reliable. This experiment is too under-specified to support the strong interpretability claim in RQ4; please provide prompt details and significance, or soften the conclusion.

minor comments (5)

[Abstract and Section 1] The first sentence of the abstract contains a grammar error ('Largelanguagemodel(LLM)have'), and elsewhere 'sarcasm' and 'irony' are used interchangeably when describing datasets such as IAC-V1 and MuSTARD, which are sarcasm benchmarks. Please clarify the intended task relationship and label conversion.
[Figure 3] The figure title contains a typo ('experimnet') and the ablation figure provides no numerical values or error bars; reporting the exact ablation numbers in a table would make the component-necessity claims verifiable.
[Section 4.6, Table 4] The inference-efficiency comparison uses a '400-sample subset' without specifying its composition or sampling procedure; this limits the reproducibility and interpretability of the efficiency results.
[Table 2] Labeling is inconsistent: 'GPT-4o+Goc' in Table 2 should likely be 'GPT-4o+GoC' to match Section 4.1, and 'GPT-4o+Boc' should be 'GPT-4o+BoC.'
[Overall] The paper does not include a code or artifact release statement. Given the API-dependent multi-agent pipeline, an artifact with prompts and split definitions would substantially improve reproducibility.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: CAF-I's SOTA claim is an empirical benchmark comparison against external labels, with no fitted-parameter or self-citation chain that forces the result.

full rationale

The paper makes no formal derivation that could reduce to its inputs. CAF-I is a hand-specified LLM prompting workflow; Eqs. (1)-(12) are procedural descriptions of role prompts and control rules, not mathematical entailments. The central SOTA claim is evaluated on four external benchmark datasets against external labels, and the baseline numbers are cited from external prior work rather than produced by fitting CAF-I's own parameters. The only self-referential component is the auxiliary explainability experiment in Section 4.5, where CAF-I's intermediate explanations are concatenated into a GPT-3.5 Turbo prompt and improve Macro-F1 from 68.4% to 70.2%. That is a sanity check of explanation utility, not a load-bearing part of the SOTA claim. The concern that the SemEval-2018 row reports 183 samples while the canonical test set has roughly 784 tweets is a dataset-alignment and baseline-comparability question, hence a correctness risk rather than circularity: if the baselines used a different split, CAF-I's numbers would be wrong or incomparable, not equivalent to its inputs by construction. No step exhibits self-definition, fitted-input-as-prediction, load-bearing self-citation, imported uniqueness, or ansatz-by-citation. Score 0.

Assumptions & free parameters 1 free parameters · 4 assumptions · 5 invented entities

The central claim rests on the reliability of LLM backbones for the described reasoning tasks, on the correctness of the external benchmark labels and splits, and on the accuracy of baselines sourced from prior literature. The only hand-chosen rule is the refinement trigger; no continuous parameters are fitted to the test data.

free parameters (1)

Refinement trigger rule = Rneeded = (CRE == Low) OR (Fcontra == Yes), with CRE and Fcontra set by the evaluator LLM
Hand-set rule in Eq. (12) that controls whether one refinement loop runs. Not fitted to data, but it is a design choice with no sensitivity analysis; changing it could shift scores.

assumptions (4)

domain assumption The benchmark datasets (IAC-V1, IAC-V2, MuSTARD, SemEval-2018) provide correct ground-truth labels for irony and are used in their standard forms.
Section 4.1, Table 1. The paper does not describe any relabeling; accuracy of all comparisons depends on these labels.
domain assumption The reported baseline results from cited works were produced on the same data splits and evaluation protocol.
Section 4.1, 'Results are sourced from literature with aligned evaluation settings.' The SemEval size mismatch (183 vs standard) makes this questionable.
domain assumption LLM backbones (GPT-4o, Qwen 2-7B, Llama 3-8B) can reliably perform the described context, semantic, and rhetorical analyses in a zero-shot setting.
All agent workflows (Eqs. 1-9) depend on LLM reliability; no checks for LLM failures are reported.
domain assumption The external search API, when used by the Context Agent, returns relevant documents that improve irony judgments.
Section 3.3 Context Agent: D = retrieval(q); the benefit is assumed, not measured separately.

invented entities (5)

Context Agent (CA)
purpose: Analyzes internal context (entities E, relations R, theme T) and optional external knowledge to produce irony judgment y1.
Introduced in this paper; its individual contribution is only assessed via ablation, and no evidence outside this paper attests to a standalone 'context agent'.
Semantic Agent (SA)
purpose: Performs literal vs. implied meaning parsing and emotion-consistency checks to produce judgment y2.
Introduced in this paper; its individual contribution is only assessed via ablation, and no evidence outside this paper attests to a standalone 'semantic agent'.
Rhetoric Agent (RA)
purpose: Detects rhetorical devices and functions to produce judgment y3.
Introduced in this paper; its individual contribution is only assessed via ablation, and no evidence outside this paper attests to a standalone 'rhetoric agent'.
Decision Agent (DA)
purpose: Aggregates agent judgments via consensus, majority, or LLM arbitration.
Introduced in this paper; its aggregation behavior is a design choice with no external validation.
Refinement Evaluator Agent (REAgent)
purpose: Triggers a single refinement loop by generating feedback triplets (fCA, fSA, fRA) based on confidence and contradictions.
The most novel component; its benefit is shown only through ablation within this paper.

how reviews work

0 comments

Cite this review

Pith. "Pith review of CAF-I: A Collaborative Multi-Agent Framework for Enhanced Irony Detection with Large Language Models." pith.science (2026). https://pith.science/paper/SUJ42WC2

@misc{pith2026250608430,
  author       = {Pith},
  title        = {Pith review of: CAF-I: A Collaborative Multi-Agent Framework for Enhanced Irony Detection with Large Language Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SUJ42WC2}},
  note         = {Machine review of arXiv:2506.08430}
}

read the original abstract

Large language model (LLM) have become mainstream methods in the field of sarcasm detection. However, existing LLM methods face challenges in irony detection, including: 1. single-perspective limitations, 2. insufficient comprehensive understanding, and 3. lack of interpretability. This paper introduces the Collaborative Agent Framework for Irony (CAF-I), an LLM-driven multi-agent system designed to overcome these issues. CAF-I employs specialized agents for Context, Semantics, and Rhetoric, which perform multidimensional analysis and engage in interactive collaborative optimization. A Decision Agent then consolidates these perspectives, with a Refinement Evaluator Agent providing conditional feedback for optimization. Experiments on benchmark datasets establish CAF-I's state-of-the-art zero-shot performance. Achieving SOTA on the vast majority of metrics, CAF-I reaches an average Macro-F1 of 76.31, a 4.98 absolute improvement over the strongest prior baseline. This success is attained by its effective simulation of human-like multi-perspective analysis, enhancing detection accuracy and interpretability.

Figures

Figures reproduced from arXiv: 2506.08430 by the authors.

**Figure 1.** Examples illustrating three primary challenges for LLM in sarcasm detection: single perspective limitation, insufficient comprehensive understanding, and lack of interpretability. Motivated by Chain-of-Thought (CoT) prompting [26] and multi-agent cooperation principles [12], this paper introduces CAF-I, a novel LLM-driven multiagent framework to address existing limitations in irony detection by simulating human-… view at source ↗

**Figure 2.** The Overall architecture of CAF-I. The framework includes three specialized agents, Context, Semantic, and Rhetoric, which conduct multi-round analysis and optimization of the input text. The Decision Agent integrates the preliminary classification, and the Evaluator Agent determines if a round of feedback optimization is needed. 3.3 Agent Design Context Agent [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗

**Figure 3.** The ablation experimnet result of CA, SA, RA and REAgent [PITH_FULL_IMAGE:figures/full_fig_p011_3.png] view at source ↗

Figures from the paper (1 more)

**Figure 4.** Figure 4: Illustration of the CAF-I framework’s step-by-step reasoning and refinement process on a input example. Case Study Initially, an incorrect "Sarcasm" judgment arose because the Semantic and Rhetoric Agents over-interpreted certain phrases, despite the Context Agent fi…

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

33 extracted references · 18 canonical work pages

[1]

NTUA-SLP at SemEval-2018 Task 3: Tracking Ironic Tweets using Ensembles of Word and Character Level Attentive RNNs

Baziotis, C., Athanasiou, N., Papalampidi, P., Kolovou, A., Paraskevopoulos, G., Ellinas, N., Potamianos, A.: Ntua-slp at semeval-2018 task 3: Tracking ironic tweets using ensembles of word and character level attentive rnns. arXiv preprint arXiv:1804.06659 (2018)

work page Pith review arXiv 2018
[2]

In: Proceed- ings of the AAAI Conference on Artificial Intelligence

Besta, M., Blach, N., Kubicek, A., Gerstenberger, R., Podstawski, M., Giani- nazzi, L., Gajda, J., Lehmann, T., Niewiadomski, H., Nyczyk, P., et al.: Graph of thoughts: Solving elaborate problems with large language models. In: Proceed- ings of the AAAI Conference on Artificial Intelligence. vol. 38, pp. 17682–17690 (2024)

work page 2024
[3]

University of Chicago Press (1974)

Booth, W.C.: A rhetoric of irony. University of Chicago Press (1974)

work page 1974
[4]

arXiv preprint arXiv:1906.01815 (2019)

Castro, S., Hazarika, D., Pérez-Rosas, V., Zimmermann, R., Mihalcea, R., Poria, S.: Towards multimodal sarcasm detection (an _obviously_ perfect paper). arXiv preprint arXiv:1906.01815 (2019)

arXiv 2019
[5]

In: Proceedings of the fourteenth conference on computational natural language learning

Davidov, D., Tsur, O., Rappoport, A.: Semi-supervised recognition of sarcasm in twitter and amazon. In: Proceedings of the fourteenth conference on computational natural language learning. pp. 107–116 (2010)

work page 2010
[6]

In: Proceedings of the 2019 conference of the North American chapter of the association for computational linguistics: human language technologies, volume 1 (long and short papers)

Devlin, J., Chang, M.W., Lee, K., Toutanova, K.: Bert: Pre-training of deep bidi- rectional transformers for language understanding. In: Proceedings of the 2019 conference of the North American chapter of the association for computational linguistics: human language technologies, volume 1 (long and short papers). pp. 4171–4186 (2019)

2019
[7]

In: Forty-first International Conference on Machine Learning (2023)

Du, Y., Li, S., Torralba, A., Tenenbaum, J.B., Mordatch, I.: Improving factual- ity and reasoning in language models through multiagent debate. In: Forty-first International Conference on Machine Learning (2023)

2023
[8]

In: Proceedings of the 7th workshop on computational approaches to subjectivity, sentiment and social media analysis

Ghosh, A., Veale, T.: Fracking sarcasm using neural network. In: Proceedings of the 7th workshop on computational approaches to subjectivity, sentiment and social media analysis. pp. 161–169 (2016)

work page 2016

Show all 33 references

[9]

arXiv preprint arXiv:2209.06170 (2022)

Kader, F.B., Nujat, N.H., Sogir, T.B., Kabir, M., Mahmud, H., Hasan, K.: Com- putational sarcasm analysis on social media: a systematic review. arXiv preprint arXiv:2209.06170 (2022)

2022 arXiv
[10]

Advances in Neural Information Processing Systems36, 51991–52008 (2023)

Li, G., Hammoud, H., Itani, H., Khizbullin, D., Ghanem, B.: Camel: Communica- tive agents for" mind" exploration of large language model society. Advances in Neural Information Processing Systems36, 51991–52008 (2023)

2023
[11]

In: Proceedings of the 60th Annual Meeting ofthe Associationfor Computational Linguistics (Volume 1: Long Papers)

Liang, B., Lou, C., Li, X., Yang, M., Gui, L., He, Y., Pei, W., Xu, R.: Multi-modal sarcasm detection via cross-modal graph convolutional network. In: Proceedings of the 60th Annual Meeting ofthe Associationfor Computational Linguistics (Volume 1: Long Papers). pp. 1767–1777 (2022)

2022
[12]

arXiv preprint arXiv:2308.03688 (2023)

Liu,X.,Yu,H.,Zhang,H.,Xu,Y.,Lei,X.,Lai,H.,Gu,Y.,Ding,H.,Men,K.,Yang, K., et al.: Agentbench: Evaluating llms as agents. arXiv preprint arXiv:2308.03688 (2023)

2023 arXiv
[13]

arXiv preprint arXiv:1907.11692 (2019)

Liu, Y., Ott, M., Goyal, N., Du, J., Joshi, M., Chen, D., Levy, O., Lewis, M., Zettlemoyer, L., Stoyanov, V.: Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692 (2019)

2019 arXiv
[14]

arXiv preprint arXiv:2109.03587 (2021)

Liu, Y., Wang, Y., Sun, A., Meng, X., Li, J., Guo, J.: A dual-channel frame- work for sarcasm recognition by detecting sentiment conflict. arXiv preprint arXiv:2109.03587 (2021)

2021 arXiv
[15]

apparently bootstrapping improves the perfor- mance of sarcasm and nastiness classifiers for online dialogue

Lukin, S., Walker, M.: Really? well. apparently bootstrapping improves the perfor- mance of sarcasm and nastiness classifiers for online dialogue. NAACL 2013 p. 30 (2013) CAF-I: Collaborative Multi-Agent Irony Detection Framework 15

2013
[16]

arXiv preprint arXiv:1301.3781 (2013)

Mikolov, T., Chen, K., Corrado, G., Dean, J.: Efficient estimation of word repre- sentations in vector space. arXiv preprint arXiv:1301.3781 (2013)

2013 arXiv
[17]

arXiv preprint arXiv:1709.05404 (2017)

Oraby, S., Harrison, V., Reed, L., Hernandez, E., Riloff, E., Walker, M.: Creat- ing and characterizing a diverse corpus of sarcasm in dialogue. arXiv preprint arXiv:1709.05404 (2017)

2017 arXiv
[18]

In: ECAI 2020, pp

Pan, H., Lin, Z., Fu, P., Wang, W.: Modeling the incongruity between sentence snippets for sarcasm detection. In: ECAI 2020, pp. 2132–2139. IOS Press (2020)

2020
[19]

In: Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP)

Pennington, J., Socher, R., Manning, C.D.: Glove: Global vectors for word repre- sentation. In: Proceedings of the 2014 conference on empirical methods in natural language processing (EMNLP). pp. 1532–1543 (2014)

2014
[20]

arXiv preprint arXiv:1610.08815 (2016)

Poria, S., Cambria, E., Hazarika, D., Vij, P.: A deeper look into sarcastic tweets using deep convolutional neural networks. arXiv preprint arXiv:1610.08815 (2016)

2016 arXiv
[21]

arXiv preprint arXiv:2412.12808 (2024)

Qiu, Z., Yu, J., Zhang, Y., Lai, H., Rao, Y., Su, Q., Yin, J.: Detecting emotional in- congruity of sarcasm by commonsense reasoning. arXiv preprint arXiv:2412.12808 (2024)

2024 arXiv
[22]

Language resources and evaluation47, 239–268 (2013)

Reyes, A., Rosso, P., Veale, T.: A multidimensional approach for detecting irony in twitter. Language resources and evaluation47, 239–268 (2013)

2013
[23]

arXiv preprint arXiv:1805.02856 (2018)

Tay, Y., Tuan, L.A., Hui, S.C., Su, J.: Reasoning with sarcasm by reading in- between. arXiv preprint arXiv:1805.02856 (2018)

2018 arXiv
[24]

In: Proceedings of the 12th international workshop on semantic evaluation

Van Hee, C., Lefever, E., Hoste, V.: Semeval-2018 task 3: Irony detection in english tweets. In: Proceedings of the 12th international workshop on semantic evaluation. pp. 39–50 (2018)

2018
[25]

Wallace, B.C., Charniak, E., et al.: Sparse, contextually informed models for irony detection: Exploiting user communities, entities and sentiment. In: Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Confe...

2015
[26]

Advances in neural information processing systems35, 24824–24837 (2022)

Wei, J., Wang, X., Schuurmans, D., Bosma, M., Xia, F., Chi, E., Le, Q.V., Zhou, D., et al.: Chain-of-thought prompting elicits reasoning in large language models. Advances in neural information processing systems35, 24824–24837 (2022)

2022
[27]

arXiv preprint arXiv:2308.08155 (2023)

Wu, Q., Bansal, G., Zhang, J., Wu, Y., Li, B., Zhu, E., Jiang, L., Zhang, X., Zhang, S., Liu, J., et al.: Autogen: Enabling next-gen llm applications via multi- agent conversation. arXiv preprint arXiv:2308.08155 (2023)

2023 arXiv
[28]

arXiv preprint arXiv:2306.01337 (2023)

Wu, Y., Jia, F., Zhang, S., Li, H., Zhu, E., Wang, Y., Lee, Y.T., Peng, R., Wu, Q., Wang, C.: Mathchat: Converse to tackle challenging math problems with llm agents. arXiv preprint arXiv:2306.01337 (2023)

2023 arXiv
[29]

In: Proceedings of the 2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation (LREC-COLING 2024)

Xue, H., Xu, L., Tong, Y., Li, R., Lin, J., Jiang, D.: Breakthrough from nuance and inconsistency: Enhancing multimodal sarcasm detection with context-aware self- attention fusion and word weight calculation. In: Proceedings of the 2024 Joint International Conference on Comput...

2024
[30]

Yao, B., Zhang, Y., Li, Q., Qin, J.: Is sarcasm detection a step-by-step reasoning process in large language models? arXiv preprint arXiv:2407.12725 (2024)

2024 arXiv
[31]

In: Proceedings of COLING 2016, the 26th International Conference on Compu- tational Linguistics: technical papers

Zhang, M., Zhang, Y., Fu, G.: Tweet sarcasm detection using deep neural network. In: Proceedings of COLING 2016, the 26th International Conference on Compu- tational Linguistics: technical papers. pp. 2449–2460 (2016)

2016
[32]

arXiv preprint arXiv:2408.11319 (2024)

Zhang, Y., Zou, C., Lian, Z., Tiwari, P., Qin, J.: Sarcasmbench: Towards evaluating large language models on sarcasm understanding. arXiv preprint arXiv:2408.11319 (2024)

2024 arXiv
[33]

arXiv preprint arXiv:2210.03493 (2022)

Zhang, Z., Zhang, A., Li, M., Smola, A.: Automatic chain of thought prompting in large language models. arXiv preprint arXiv:2210.03493 (2022)

2022 arXiv

Pith tools

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

[1] [1]

NTUA-SLP at SemEval-2018 Task 3: Tracking Ironic Tweets using Ensembles of Word and Character Level Attentive RNNs

Baziotis, C., Athanasiou, N., Papalampidi, P., Kolovou, A., Paraskevopoulos, G., Ellinas, N., Potamianos, A.: Ntua-slp at semeval-2018 task 3: Tracking ironic tweets using ensembles of word and character level attentive rnns. arXiv preprint arXiv:1804.06659 (2018)

work page Pith review arXiv 2018

[2] [2]

In: Proceed- ings of the AAAI Conference on Artificial Intelligence

Besta, M., Blach, N., Kubicek, A., Gerstenberger, R., Podstawski, M., Giani- nazzi, L., Gajda, J., Lehmann, T., Niewiadomski, H., Nyczyk, P., et al.: Graph of thoughts: Solving elaborate problems with large language models. In: Proceed- ings of the AAAI Conference on Artificial Intelligence. vol. 38, pp. 17682–17690 (2024)

work page 2024

[3] [3]

University of Chicago Press (1974)

Booth, W.C.: A rhetoric of irony. University of Chicago Press (1974)

work page 1974

[4] [4]

arXiv preprint arXiv:1906.01815 (2019)

Castro, S., Hazarika, D., Pérez-Rosas, V., Zimmermann, R., Mihalcea, R., Poria, S.: Towards multimodal sarcasm detection (an _obviously_ perfect paper). arXiv preprint arXiv:1906.01815 (2019)

arXiv 2019

[5] [5]

In: Proceedings of the fourteenth conference on computational natural language learning

Davidov, D., Tsur, O., Rappoport, A.: Semi-supervised recognition of sarcasm in twitter and amazon. In: Proceedings of the fourteenth conference on computational natural language learning. pp. 107–116 (2010)

work page 2010

[6] [6]

In: Proceedings of the 2019 conference of the North American chapter of the association for computational linguistics: human language technologies, volume 1 (long and short papers)

Devlin, J., Chang, M.W., Lee, K., Toutanova, K.: Bert: Pre-training of deep bidi- rectional transformers for language understanding. In: Proceedings of the 2019 conference of the North American chapter of the association for computational linguistics: human language technologies, volume 1 (long and short papers). pp. 4171–4186 (2019)

2019

[7] [7]

In: Forty-first International Conference on Machine Learning (2023)

Du, Y., Li, S., Torralba, A., Tenenbaum, J.B., Mordatch, I.: Improving factual- ity and reasoning in language models through multiagent debate. In: Forty-first International Conference on Machine Learning (2023)

2023

[8] [8]

In: Proceedings of the 7th workshop on computational approaches to subjectivity, sentiment and social media analysis

Ghosh, A., Veale, T.: Fracking sarcasm using neural network. In: Proceedings of the 7th workshop on computational approaches to subjectivity, sentiment and social media analysis. pp. 161–169 (2016)

work page 2016