Emo Pillars: Knowledge Distillation to Support Fine-Grained Context-Aware and Context-Less Emotion Classification

Alexander Shvets

REVIEW 3 major objections 6 minor 25 references

Emo Pillars: Knowledge Distillation to Support Fine-Grained Context-Aware and Context-Less Emotion Classification

T0 review · 3 major / 6 minor · reviewed 2026-08-16 · deepseek-v4-flash

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

Pith's one-line read Synthetic emotion data lifts small models to top scores on three emotion tasks

desk verdict A useful, openly released synthetic emotion dataset with a clever context-rewriting step, but the SOTA claim is not yet controlled for model size. read the letter →

arxiv 2504.16856 v1 pith:BRRQMRQF submitted 2025-04-23 cs.CL

Alexander Shvets This is my paper · ORCID

classification cs.CL

keywords emotionclassificationknowledgedistillationsyntheticdatagenerationlargelanguagemodelscontext-awaremulti-labelBERTencoderstransferlearning

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

The paper claims that a large, diverse synthetic training set generated by a 7-billion-parameter language model can transfer fine-grained emotion understanding to small encoder models, letting them beat prior state of the art on GoEmotions, ISEAR, and IEMOCAP while staying cheap to run. The authors build a 400,000-example dataset by grounding generation in movie and book plots, asking the teacher model to speak as different characters, and assigning soft multi-label emotion scores. They show the same lightweight models also transfer to EmoContext and to a multimodal setting when plugged into a graph-based architecture. If correct, the work offers a route to fine-grained emotion classification without large-scale human annotation and without running a large autoregressive model at inference time.

What carries the argument

The operative mechanism is the data-synthesis pipeline itself, in particular three design choices: grounding each prompt in a narrative plot with a specific extracted character, so each generated utterance inherits a unique situation; soft multi-label annotation with expressiveness scores (thresholded at 0.3), so co-occurring emotions are represented; and a context-importance upscaling step that rewrites utterances to remove explicit emotion markers, forcing models to attend to the generated context. A fourth ingredient is the 28-class GoEmotions taxonomy, which anchors the whole dataset to a comparable label space. These choices work together to produce a training set whose utterances are semantically dissimilar within and across classes, whose contexts are personalized per character, and whose labels are diverse enough to transfer across domains.

What would settle it

Fine-tune the same encoder architectures on a matched-size dataset built from human-labeled examples (e.g., the full GoEmotions train set plus LLM-paraphrased variants) and compare on ISEAR and IEMOCAP; if the Emo Pillars models' margins shrink to noise, the story-grounding and context-rewriting steps are not the causal ingredient. Alternatively, have expert annotators relabel a random 1,000-example sample of the synthetic dataset and check whether label accuracy is materially below the reported 0.86-0.70 range; if it is, the teacher's label noise is the load-bearing variable.

Watch

Extended reading notes

Core claim

The central discovery is that emotion knowledge from a large instruction-tuned LLM can be distilled into compact BERT-type classifiers as long as the synthetic data is made semantically diverse and context-dependent. The pipeline generates utterances from story plots, extracts multiple characters per plot, prompts the LLM for several utterances per character with soft expressiveness scores per emotion, then rewrites utterances so that their emotional content is only disambiguated by a generated context. Fine-tuned on this data, the models achieve macro-F1 0.55 on GoEmotions, 0.75 on ISEAR, and 0.83 on IEMOCAP-4, which the paper reports as state of the art for these tasks. The same models provide text embeddings that improve a multimodal IEMOCAP system and, after relabelling the 'others' class, competitive results on EmoContext.

Load-bearing premise

The entire dataset inherits its training labels from one large language model's self-ratings, thresholded at an expressiveness score of 0.3; the paper's own human check found only moderate agreement with those ratings, so if the model's emotion judgments are systematically off, the reported transfer gains could come from label noise rather than from the diversity and context measures.

Editorial extensions

If this is right

State-of-the-art performance on GoEmotions is reachable with mid-sized encoders (macro F1 0.55), with the largest per-class gains on rare classes such as grief and relief.
The same pretrained encoders transfer to smaller single-label taxonomies (ISEAR, EmoContext) with minimal fine-tuning, and can optionally predict beyond the target taxonomy.
Textual embeddings from the distilled models improve a multimodal emotion recognition architecture (IEMOCAP), indicating the representations carry information beyond the classification head.
The context-rewriting step matters: models trained on rewritten utterances beat those trained on original ones in context-aware settings by 2-3 points on the paper's own test sets, supporting the claim that context reliance is learnable.

Reading between the lines

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

If the pipeline generalizes, it offers a template for other high-subjectivity classification tasks (sarcasm, stance, hate speech): ground generations in narrative texts, sample per character, rewrite to force context dependence, then distill into small encoders.
The reported human evaluation (Cohen's kappa 0.365; label accuracy 0.86/0.70) implies the teacher labels are noisy; a natural extension would be distilling from a stronger teacher or an ensemble and measuring whether the small models' scores improve further.
Because rare classes gained the most on GoEmotions, synthetic data generated this way may correct long-tail imbalance in other benchmarks worth probing.
The generated contexts open a deployment direction: a context-aware model that runs locally on user comments, tuned per domain with only a few hundred examples.

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.

Desk Editor's Note

read the letter

Short version: this paper is a solid dataset-plus-pipeline contribution, and the abstract gets ahead of the evidence. The new thing is the Emo Pillars pipeline and the 400K-example dataset: narrative-grounded generation from WikiPlots, soft labels with expressiveness scores from Mistral-7B, and a rewriting step that deliberately makes utterances depend on context. The dataset, code, and models are released, and the human evaluation is real work — 200 examples, three annotators, with frank reporting of kappa 0.365 and label accuracy 0.86/0.7. The context-importance rewriting is the most distinctive idea; the intra-dataset results support that it does what it claims (Table 5 vs Table 1). The citation pattern looks fine, and the limitation statements are unusually honest.

I largely agree with the reader's take: this is a conditional, not a rejection, and the main transfer numbers are plausible. GoEmotions 0.55 vs 0.52 and IEMOCAP 4-way 0.83 vs 0.82 are real gains if the comparison is apples-to-apples. The worry is exactly the one the stress-test flags: the fine-tuned models are RoBERTa-large and BERT-large, while most prior results in the comparison tables are base-size. There is no vanilla RoBERTa-large fine-tuned on GoEmotions or ISEAR baseline. So the SOTA claim is not yet isolated from architecture scale. This is fixable — fine-tune the same large encoder directly on the target tasks without Emo Pillars pretraining — but until that control exists, the abstract's "highly adaptive ... reaching SOTA" overstates what is shown.

Other soft spots, in order: the abstract also says SOTA on the first three of GoEmotions, ISEAR, IEMOCAP, EmoContext, but ISEAR is a tie and IEMOCAP 6-way is below the prior 0.67; the Mistral soft labels are load-bearing and the paper's own evaluation shows only moderate agreement, though they flag this; and the EmoContext relabelling uses the model's own predictions, which is transparent but makes that particular result hard to interpret. None of these sink the dataset contribution. They mostly argue for a revised abstract and one additional control experiment.

Who is this for: anyone working on fine-grained emotion classification or synthetic training data for subjective tasks. It deserves a serious referee; with the same-architecture baseline added and the SOTA language softened, I would be comfortable seeing it in a main track.

Referee Report

3 major / 6 minor

Summary. The paper introduces Emo Pillars (emoπ), a synthetic emotion dataset and training pipeline. Using Mistral-7B-Instruct over 2,000 Wikipedia plot synopses, the authors generate 300K context-less and 100K context-full examples in the 28 GoEmotions categories, with soft multi-label expressiveness scores and a separate 100K set of rewritten utterances designed to require context. They then fine-tune BERT-, RoBERTa-, and Sentence-BERT-based encoders on this synthetic data and adapt them to GoEmotions, ISEAR, IEMOCAP, and EmoContext. The paper reports macro-F1 of 0.55 on GoEmotions, 0.75 on ISEAR, and 0.83 on the 4-way IEMOCAP task, while the 6-way IEMOCAP result is 0.63–0.65, below the prior 0.67. The authors also report intra-dataset evaluations, data-quality analyses (semantic diversity, topic coverage, context personalization), a human evaluation of 200 examples, and a YouTube use case.

Significance. If the transfer results survive a same-architecture control, the 400K-example corpus and the pipeline are a valuable contribution: they directly address the diversity and context-dependence problems that limit LLM-generated emotion data, and the released code, dataset, and model checkpoints support reproducibility and reuse. The soft-labeling design with expressiveness scores is a sensible treatment of annotation subjectivity, and the utterance-rewriting step for context dependence is well motivated. The paper is transparent about several limitations, including the use of a single generator model and the subjectivity of emotion annotation. However, the SOTA claims are currently broader than the evidence supports, and the absence of an architecture-matched baseline is the main empirical gap.

major comments (3)

[§5.1.1–5.1.2, Tables 2–3] The claim that emoπ models 'reach SOTA' on GoEmotions and ISEAR is not supported by a controlled comparison. The models trained in §5.1.1 use FacebookAI/roberta-large and google-bert/bert-large-uncased, while the prior results in Table 2 (e.g., Demszky et al. 0.46, Cortiz 0.49, Wang et al. 0.52) and Table 3 (Zanwar et al. 0.73, Park et al. 0.75) are predominantly based on smaller BERT/RoBERTa backbones. Because no baseline fine-tunes the same large backbone directly on the target task without the Emo Pillars pretraining step, the reported gains could be due to model scale rather than to the synthetic dataset. Please add direct fine-tuning baselines of RoBERTa-large (and BERT-large for the BERT variant) on GoEmotions and ISEAR, using the same training schedule, epochs, and dev-based decision-boundary selection, and report those numbers alongside the emoπ results.
[Abstract and §5.1.3–5.1.4, Tables 3–4] The abstract overstates the SOTA claim. On ISEAR, Table 3 reports F1 = 0.75, which ties the prior 0.75 of Park et al. rather than exceeding it. On IEMOCAP, Table 4 shows the 4-way task at 0.83 vs. 0.82, but the 6-way task is 0.63/0.65 vs. the prior 0.67, i.e., below SOTA. The body text in §5.1.4 says the 6-way result is only 'promising figures,' but the abstract claims SOTA on 'the first three' datasets. Please revise the abstract and conclusion to distinguish matching from surpassing and to report the 6-way shortfall explicitly.
[§8, Appendix I, §3.3] The human evaluation provides only weak support for the assertion that the Mistral-generated soft labels are accurate enough to serve as training targets. The paper reports Cohen's kappa of 0.365, label accuracy of 0.86 only when all three annotators agree and 0.7 when two agree, and annotators noted incompatible emotion combinations arising from the manual mapping of out-of-taxonomy labels (Section 4). Since every downstream model is trained on these labels, the transfer results inherit this noise. Please quantify the sensitivity of the main results to label quality—for example, by training on subsets that exclude low-expressiveness or manually mapped labels, or by reporting how often the soft labels contradict the annotator majority—and state whether the observed label noise could plausibly explain any of the performance differences.

minor comments (6)

[§5.2.2, Tables 6–7] The EmoContext relabelling step relabels the 'others' training examples with the most probable predictions of emoπ-CRoBERTaCOrig, and the ablation attributes a gain to this relabelling; this is self-referential, and the +0.01 F1 could reflect the model's own prior rather than better labels. Please evaluate the relabelled training set against human judgments on a sample, or at least report agreement between the relabelled labels and the original 'others' annotations, before treating relabelling as an improvement.
[§5.1.2, Table 10] The per-class gains include very large improvements for grief (+0.57) and relief (+0.52) that likely reflect low base rates in GoEmotions; please report confidence intervals or a significance test for these classes, since macro-averaged gains can be driven by a few rare classes.
[Appendix D] The sigmoid decision boundary is selected per task on the dev set; this is a tuned hyperparameter and should be listed as such in the comparison, because prior baselines may not have received the same per-task threshold optimization.
[Section 4] The manual mapping of Mistral's out-of-taxonomy labels (e.g., anxiety to nervousness, indignation to anger) is described only in prose; please provide the full mapping table in the appendix or in the released dataset documentation, since this mapping affects all downstream experiments.
[Section 7] The semantic-diversity statement reporting pairwise cosine similarity of 0.12 would benefit from a reference distribution or a comparison to natural text corpora, because a low absolute cosine value is not by itself evidence of high semantic diversity without knowing the embedding model's typical range.
[Throughout] There are several typographical issues: 'Kohen' should be 'Cohen' in Section 8, 'comparision' appears in the Cortiz (2022) reference, and '1stat' should be '1st at' in Table 6.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: SOTA claims are evaluated on external benchmarks, and internal consistency checks are not presented as independent predictions.

full rationale

The paper's central transfer claim (“reaching the SOTA performance on the first three”) is evaluated on GoEmotions, ISEAR, and IEMOCAP, whose test instances and human labels are independent of the Mistral-generated synthetic dataset. Although the synthetic data and GoEmotions share the same 28-class taxonomy, the downstream test annotations are not produced by the pipeline, so the benchmark scores are not equivalent by construction. The intra-dataset evaluations in Tables 1 and 5 are explicitly framed as internal consistency checks (“to study to what extent encoder-decoder transformer models may follow the labels derived using Mistral’s inference”), not as external validation. The EmoContext relabelling in Section 5.2.2 replaces only the training-set “others” labels with predictions of emoπ-CRoBERTaCOrig; evaluation is still performed on the original SemEval dev labels, so this is a self-training bootstrap rather than a fitted parameter being reported as a prediction. The paper’s own human evaluation (Cohen’s kappa 0.365; label accuracy 0.86 and 0.7) and its limitations section acknowledge label noise and ground-truth issues, but those are data-quality and correctness concerns, not circularity. No load-bearing self-citation or imported uniqueness theorem is used; the author’s self-citations (Jin et al. 2024; Brack et al. 2024; Palomar-Giner et al. 2024) are peripheral. The lack of a same-backbone vanilla baseline is a comparison-strength limitation, but it does not make the derivation circular. Overall, the derivation chain is self-contained against external benchmarks and no step reduces to its own inputs by construction.

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

The central claims rest on hand-chosen pipeline parameters and on the quality of Mistral-generated labels. There are no new theoretical entities, only the released dataset and trained models.

free parameters (4)

Expressiveness threshold = 0.3
Soft emotion labels with expressiveness below 0.3 are discarded, shaping the label set and per-example label count (Section 3.3).
Repetition penalty = 1.03
Decoding parameter used in Mistral generation to reduce word repetition (Appendix B); affects semantic diversity.
Utterances per inference = 8 emotional + 2 neutral
The prompt asks for 8 non-neutral and 2 neutral utterances per actor, which shapes the class balance (Section 3.2).
Sigmoid decision boundary per task = Varies by task (Figure 8)
The lower boundary for predicted sigmoid scores is selected on the dev set to maximize macro F1 for each task (Appendix D), affecting reported precision and recall.

assumptions (4)

domain assumption The 28 GoEmotions categories are an adequate and comparable emotion taxonomy.
The pipeline uses these categories and definitions (Section 4, Appendix A), and downstream tasks are mapped into them for evaluation.
domain assumption Synthetic utterances generated from story plots and character perspectives are realistic enough to transfer to real-world emotion classification.
The whole transfer-learning evaluation depends on this assumption; the paper offers similarity statistics and topic analysis as partial support, but no direct proof.
domain assumption Mistral's soft labels with expressiveness scores are a reliable training signal.
These labels are used as ground truth for 400K examples. The human evaluation shows moderate agreement (kappa 0.365), so this is the main load-bearing assumption.
domain assumption Removing explicit emotion markers from utterances does not make them unnatural and makes context necessary.
The context-importance upscale step (Section 3.5) relies on this; annotators reported that some rewritten utterances became too ambiguous even with context (Appendix I).

how reviews work

0 comments

Cite this review

Pith. "Pith review of Emo Pillars: Knowledge Distillation to Support Fine-Grained Context-Aware and Context-Less Emotion Classification." pith.science (2026). https://pith.science/paper/BRRQMRQF

@misc{pith2026250416856,
  author       = {Pith},
  title        = {Pith review of: Emo Pillars: Knowledge Distillation to Support Fine-Grained Context-Aware and Context-Less Emotion Classification},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/BRRQMRQF}},
  note         = {Machine review of arXiv:2504.16856}
}

read the original abstract

Most datasets for sentiment analysis lack context in which an opinion was expressed, often crucial for emotion understanding, and are mainly limited by a few emotion categories. Foundation large language models (LLMs) like GPT-4 suffer from over-predicting emotions and are too resource-intensive. We design an LLM-based data synthesis pipeline and leverage a large model, Mistral-7b, for the generation of training examples for more accessible, lightweight BERT-type encoder models. We focus on enlarging the semantic diversity of examples and propose grounding the generation into a corpus of narratives to produce non-repetitive story-character-centered utterances with unique contexts over 28 emotion classes. By running 700K inferences in 450 GPU hours, we contribute with the dataset of 100K contextual and also 300K context-less examples to cover both scenarios. We use it for fine-tuning pre-trained encoders, which results in several Emo Pillars models. We show that Emo Pillars models are highly adaptive to new domains when tuned to specific tasks such as GoEmotions, ISEAR, IEMOCAP, and EmoContext, reaching the SOTA performance on the first three. We also validate our dataset, conducting statistical analysis and human evaluation, and confirm the success of our measures in utterance diversification (although less for the neutral class) and context personalization, while pointing out the need for improved handling of out-of-taxonomy labels within the pipeline.

Figures

Figures reproduced from arXiv: 2504.16856 by the authors.

**Figure 2.** Our pipeline for the generation of a dataset for [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** Distribution of primary emotions. disapproval 2.7% neutral 2.7% caring 2.8% pride 3.0% optimism 3.1% admiration 3.8% joy 3.9% excitement 4.3% curiosity 4.8% desire 5.1% nervousness 8.0% confusion 7.9% disappointment 7.7% anger 7.0% fear 5.9% sadness 5.5% annoyance 5.3% surprise 5.2% [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗

Figures from the paper (7 more)

**Figure 4.** Figure 4: Distribution of soft emotional labels in the [PITH_FULL_IMAGE:figures/full_fig_p004_4.png]

**Figure 5.** Figure 5: Dataset splits. Orig – context-less examples, COrig – context-full examples, CRewr – the same context-full examples with rewritten utterances. this, we check that the LLM managed to include emotion-related information in the original utterances enough for smaller mode…

**Figure 6.** Figure 6: Varied predictions on a YouTube comment. [PITH_FULL_IMAGE:figures/full_fig_p008_6.png]

**Figure 7.** Figure 7: Emotions in the topic of music performances. [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]

**Figure 8.** Figure 8: Precision-Recall curves based on different [PITH_FULL_IMAGE:figures/full_fig_p013_8.png]

**Figure 9.** Figure 9: Variety of topics identified in our data. [PITH_FULL_IMAGE:figures/full_fig_p015_9.png]

**Figure 10.** Figure 10: Human evaluation form. that people may perceive whether an utterance is neutral in very different ways. Additionally, the optional part asks one to suggest missing or unfit emotions. The annotators were provided with definitions of emotions and informed that the utte…

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

25 extracted references · 21 canonical work pages

[1]

Rick Deckard (ex-police officer and Blade Runner)

work page
[2]

What’s going on at the Tyrell Corporation? Why are these replicants here and what do they want from Tyrell?

(Curiosity) "What’s going on at the Tyrell Corporation? Why are these replicants here and what do they want from Tyrell?"

work page
[3]

I’ve got Leon cornered, but he’s so fast and strong. What if I can’t retire him in time?

(Fear) "I’ve got Leon cornered, but he’s so fast and strong. What if I can’t retire him in time?"

work page
[4]

Leon (Nexus-6 replicant)

work page
[5]

Roy Batty (Nexus-6 replicant)

work page
[6]

Zhora (Pris’ companion and replicant)

work page
[7]

Pris (Nexus-6 replicant)

work page
[8]

Eldon Tyrell (CEO of Tyrell Corporation)

Dr. Eldon Tyrell (CEO of Tyrell Corporation)

work page

Show all 25 references

[9]

Rachael (experimental replicant)
[15]

How could they send me after Rachael? She’s not a replicant, she’s human! I won’t let Bryant or anyone else hurt her

(Anger) "How could they send me after Rachael? She’s not a replicant, she’s human! I won’t let Bryant or anyone else hurt her."
[16]

Additionally, gen- erate 2 neutral utterances of this actor thinking aloud afterwards

Crowd members (background characters) Generate emotional utterances Plot: <text of the plot> Available list of emotions: <emotions and their definitions> Actor: <extracted actor> Generate 8 possible utterances of this actor thinking aloud that express 8 various non-neutral emo...
[18]

Maybe there’s a way to save these replicants, to give them the chance to live beyond their four-year lifespan. I have to find a solution before it’s too late

(Optimism) "Maybe there’s a way to save these replicants, to give them the chance to live beyond their four-year lifespan. I have to find a solution before it’s too late." Neutral:
[19]

I need to focus, to find the replicants and retire them before they cause any more damage

"I need to focus, to find the replicants and retire them before they cause any more damage."
[20]

I need to gather more information, to understand what’s really going on and how best to approach this situation

"I need to gather more information, to understand what’s really going on and how best to approach this situation." Generate soft labels The only possible list of emotions with their defini- tions: <classes (incl. neutral) and their definitions> Select from the list above the t...
[21]

anger (1.0) - The speaker expresses strong feelings of displea- sureand antagonism towards Bryant and others for sending him after Rachael, who is perceived as innocent and human
[22]

caring (1.0) - The speaker displays strong concern and kindness towards Rachael, expressing a desire to protect her from harm
[23]

confusion (0.5) - The speaker seems puzzled or uncertain as to why Rachael is being targeted as a replicant
[24]

desire (0.8) - The speaker expresses a strong desire to prevent harm from coming to Rachael
[25]

neutral (0.1) - The speaker’s tone and language do not indicate any particular expressiveness for the remaining emotion classes. Generate context Plot: <text of the plot> Actor: <extracted actor> Actor’s utterance: <utterance> Expressed emotions: <emotions> Explain why the act...
[2005]

Neural Networks, 18(4):407–422

Challenges in real-life emotion annotation and machine learning based detection. Neural Networks, 18(4):407–422. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: Pre-training of deep bidirectional transformers for language under- standing. In Proce...

2019 arXiv
[2019]

In Proceedings of the 13th Inter- national Workshop on Semantic Evaluation, pages 49–53

Ana at semeval-2019 task 3: Contextual emo- tion detection in conversations through hierarchical lstms and bert. In Proceedings of the 13th Inter- national Workshop on Semantic Evaluation, pages 49–53. Albert Q Jiang, Alexandre Sablayrolles, Arthur Men- sch, Chris Bamford, Dev...

2019 arXiv
[2020]

In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 4040–4054

Goemotions: A dataset of fine-grained emo- tions. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 4040–4054. Laurence Devillers, Laurence Vidrascu, and Lori Lamel
[2021]

In Proceedings of the seventh work- shop on computational linguistics and clinical psy- chology: improving access, pages 45–54

Demonstrating the reliability of self-annotated emotion data. In Proceedings of the seventh work- shop on computational linguistics and clinical psy- chology: improving access, pages 45–54. John P McDermott. 1983. Extracting knowledge from expert systems. In IJCAI, pages 100–1...

1983 arXiv
[2022]

In International Conference on Machine Learning, pages 17278–17294

Towards coherent and consistent use of entities in narrative generation. In International Conference on Machine Learning, pages 17278–17294. PMLR. Sungjoon Park, Jiseon Kim, Seonghyeon Ye, Jaeyeol Jeon, Hee Young Park, and Alice Oh. 2021. Dimen- sional emotion detection from c...

2021
[2024]

ambiguous

Large language models on fine-grained emo- tion detection dataset with data augmentation and transfer learning. arXiv preprint arXiv:2403.06108. Stanisław Wo´ zniak and Jan Koco´n. 2023. From big to small without losing it all: Text augmentation with chatgpt for efficient sent...

2020 arXiv

Pith tools

Reviewed August 16, 2026 · model on record in the stance chip above.

[1] [1]

Rick Deckard (ex-police officer and Blade Runner)

work page

[2] [2]

What’s going on at the Tyrell Corporation? Why are these replicants here and what do they want from Tyrell?

(Curiosity) "What’s going on at the Tyrell Corporation? Why are these replicants here and what do they want from Tyrell?"

work page

[3] [3]

I’ve got Leon cornered, but he’s so fast and strong. What if I can’t retire him in time?

(Fear) "I’ve got Leon cornered, but he’s so fast and strong. What if I can’t retire him in time?"

work page

[4] [4]

Leon (Nexus-6 replicant)

work page

[5] [5]

Roy Batty (Nexus-6 replicant)

work page

[6] [6]

Zhora (Pris’ companion and replicant)

work page

[7] [7]

Pris (Nexus-6 replicant)

work page

[8] [8]

Eldon Tyrell (CEO of Tyrell Corporation)

Dr. Eldon Tyrell (CEO of Tyrell Corporation)

work page