HEART: Emotionally-driven test-time scaling of Language Models Article Swipe
YOU?
·
· 2025
· Open Access
·
· DOI: https://doi.org/10.48550/arxiv.2509.22876
Test-time scaling has shown considerable success in improving the performance of language models on complex reasoning tasks without requiring fine-tuning. However, current strategies such as self-reflection primarily focus on logical or structural refinement. They do not leverage the guiding potential of affective feedback. Inspired by psychological research showing that emotions can modulate cognitive performance, we introduce HEART--a novel framework that uses emotionally-driven prompts for iterative self-correction. HEART provides feedback on a model's incorrect response using a curated set of concise, emotionally charged phrases based on the six universal emotions categorized by Dr. Paul Ekman. By systematically varying the emotional tone of the feedback across iterations, our method guides the model to escape flawed reasoning paths and explore more promising alternatives. We evaluate our framework on challenging reasoning benchmarks including OlympiadBench, Humanity's Last Exam, and SimpleQA. Our results reveal a significant new phenomenon: when guided by an oracle verifier, this affective iteration protocol unlocks significantly deeper reasoning, leading to consistent and substantial increases in accuracy over state-of-the-art baselines with the same verifier. However, we also identify a critical bottleneck for practical deployment. In a verifier-free setting, it struggles to harness these gains consistently, highlighting as a key challenge for future work. Our findings suggest that the next frontier in machine reasoning may lie not just in refining logic, but also in understanding and leveraging the `HEART' of the models.
Related Topics
- Type
- preprint
- Language
- en
- Landing Page
- http://arxiv.org/abs/2509.22876
- https://arxiv.org/pdf/2509.22876
- OA Status
- green
- OpenAlex ID
- https://openalex.org/W4415331397
Raw OpenAlex JSON
- OpenAlex ID
-
https://openalex.org/W4415331397Canonical identifier for this work in OpenAlex
- DOI
-
https://doi.org/10.48550/arxiv.2509.22876Digital Object Identifier
- Title
-
HEART: Emotionally-driven test-time scaling of Language ModelsWork title
- Type
-
preprintOpenAlex work type
- Language
-
enPrimary language
- Publication year
-
2025Year of publication
- Publication date
-
2025-09-26Full publication date if available
- Authors
-
Gabriela Souza de Oliveira Pinto, Palash Goyal, Yong Sang Song, Subenoy Chakraborty, Zifeng Wang, Tomas Pfister, Hamid PalangiList of authors in order
- Landing page
-
https://arxiv.org/abs/2509.22876Publisher landing page
- PDF URL
-
https://arxiv.org/pdf/2509.22876Direct link to full text PDF
- Open access
-
YesWhether a free full text is available
- OA status
-
greenOpen access status per OpenAlex
- OA URL
-
https://arxiv.org/pdf/2509.22876Direct OA link when available
- Cited by
-
0Total citation count in OpenAlex
Full payload
| id | https://openalex.org/W4415331397 |
|---|---|
| doi | https://doi.org/10.48550/arxiv.2509.22876 |
| ids.doi | https://doi.org/10.48550/arxiv.2509.22876 |
| ids.openalex | https://openalex.org/W4415331397 |
| fwci | |
| type | preprint |
| title | HEART: Emotionally-driven test-time scaling of Language Models |
| biblio.issue | |
| biblio.volume | |
| biblio.last_page | |
| biblio.first_page | |
| topics[0].id | https://openalex.org/T10028 |
| topics[0].field.id | https://openalex.org/fields/17 |
| topics[0].field.display_name | Computer Science |
| topics[0].score | 0.9362000226974487 |
| topics[0].domain.id | https://openalex.org/domains/3 |
| topics[0].domain.display_name | Physical Sciences |
| topics[0].subfield.id | https://openalex.org/subfields/1702 |
| topics[0].subfield.display_name | Artificial Intelligence |
| topics[0].display_name | Topic Modeling |
| is_xpac | False |
| apc_list | |
| apc_paid | |
| language | en |
| locations[0].id | pmh:oai:arXiv.org:2509.22876 |
| locations[0].is_oa | True |
| locations[0].source.id | https://openalex.org/S4306400194 |
| locations[0].source.issn | |
| locations[0].source.type | repository |
| locations[0].source.is_oa | True |
| locations[0].source.issn_l | |
| locations[0].source.is_core | False |
| locations[0].source.is_in_doaj | False |
| locations[0].source.display_name | arXiv (Cornell University) |
| locations[0].source.host_organization | https://openalex.org/I205783295 |
| locations[0].source.host_organization_name | Cornell University |
| locations[0].source.host_organization_lineage | https://openalex.org/I205783295 |
| locations[0].license | |
| locations[0].pdf_url | https://arxiv.org/pdf/2509.22876 |
| locations[0].version | submittedVersion |
| locations[0].raw_type | text |
| locations[0].license_id | |
| locations[0].is_accepted | False |
| locations[0].is_published | False |
| locations[0].raw_source_name | |
| locations[0].landing_page_url | http://arxiv.org/abs/2509.22876 |
| locations[1].id | doi:10.48550/arxiv.2509.22876 |
| locations[1].is_oa | True |
| locations[1].source.id | https://openalex.org/S4306400194 |
| locations[1].source.issn | |
| locations[1].source.type | repository |
| locations[1].source.is_oa | True |
| locations[1].source.issn_l | |
| locations[1].source.is_core | False |
| locations[1].source.is_in_doaj | False |
| locations[1].source.display_name | arXiv (Cornell University) |
| locations[1].source.host_organization | https://openalex.org/I205783295 |
| locations[1].source.host_organization_name | Cornell University |
| locations[1].source.host_organization_lineage | https://openalex.org/I205783295 |
| locations[1].license | cc-by |
| locations[1].pdf_url | |
| locations[1].version | |
| locations[1].raw_type | article |
| locations[1].license_id | https://openalex.org/licenses/cc-by |
| locations[1].is_accepted | False |
| locations[1].is_published | |
| locations[1].raw_source_name | |
| locations[1].landing_page_url | https://doi.org/10.48550/arxiv.2509.22876 |
| indexed_in | arxiv, datacite |
| authorships[0].author.id | https://openalex.org/A5071073494 |
| authorships[0].author.orcid | |
| authorships[0].author.display_name | Gabriela Souza de Oliveira Pinto |
| authorships[0].author_position | first |
| authorships[0].raw_author_name | Pinto, Gabriela |
| authorships[0].is_corresponding | False |
| authorships[1].author.id | https://openalex.org/A5011797382 |
| authorships[1].author.orcid | https://orcid.org/0000-0003-2455-2160 |
| authorships[1].author.display_name | Palash Goyal |
| authorships[1].author_position | middle |
| authorships[1].raw_author_name | Goyal, Palash |
| authorships[1].is_corresponding | False |
| authorships[2].author.id | https://openalex.org/A5004406086 |
| authorships[2].author.orcid | https://orcid.org/0000-0001-7115-4021 |
| authorships[2].author.display_name | Yong Sang Song |
| authorships[2].author_position | middle |
| authorships[2].raw_author_name | Song, Yiwen |
| authorships[2].is_corresponding | False |
| authorships[3].author.id | https://openalex.org/A5022000851 |
| authorships[3].author.orcid | https://orcid.org/0000-0002-3881-8022 |
| authorships[3].author.display_name | Subenoy Chakraborty |
| authorships[3].author_position | middle |
| authorships[3].raw_author_name | Chakraborty, Souradip |
| authorships[3].is_corresponding | False |
| authorships[4].author.id | https://openalex.org/A5067779271 |
| authorships[4].author.orcid | https://orcid.org/0000-0002-3026-9970 |
| authorships[4].author.display_name | Zifeng Wang |
| authorships[4].author_position | middle |
| authorships[4].raw_author_name | Wang, Zifeng |
| authorships[4].is_corresponding | False |
| authorships[5].author.id | https://openalex.org/A5101265241 |
| authorships[5].author.orcid | |
| authorships[5].author.display_name | Tomas Pfister |
| authorships[5].author_position | middle |
| authorships[5].raw_author_name | Pfister, Tomas |
| authorships[5].is_corresponding | False |
| authorships[6].author.id | https://openalex.org/A5033846851 |
| authorships[6].author.orcid | https://orcid.org/0000-0003-2912-4579 |
| authorships[6].author.display_name | Hamid Palangi |
| authorships[6].author_position | last |
| authorships[6].raw_author_name | Palangi, Hamid |
| authorships[6].is_corresponding | False |
| has_content.pdf | False |
| has_content.grobid_xml | False |
| is_paratext | False |
| open_access.is_oa | True |
| open_access.oa_url | https://arxiv.org/pdf/2509.22876 |
| open_access.oa_status | green |
| open_access.any_repository_has_fulltext | False |
| created_date | 2025-10-19T00:00:00 |
| display_name | HEART: Emotionally-driven test-time scaling of Language Models |
| has_fulltext | False |
| is_retracted | False |
| updated_date | 2025-11-06T06:51:31.235846 |
| primary_topic.id | https://openalex.org/T10028 |
| primary_topic.field.id | https://openalex.org/fields/17 |
| primary_topic.field.display_name | Computer Science |
| primary_topic.score | 0.9362000226974487 |
| primary_topic.domain.id | https://openalex.org/domains/3 |
| primary_topic.domain.display_name | Physical Sciences |
| primary_topic.subfield.id | https://openalex.org/subfields/1702 |
| primary_topic.subfield.display_name | Artificial Intelligence |
| primary_topic.display_name | Topic Modeling |
| cited_by_count | 0 |
| locations_count | 2 |
| best_oa_location.id | pmh:oai:arXiv.org:2509.22876 |
| best_oa_location.is_oa | True |
| best_oa_location.source.id | https://openalex.org/S4306400194 |
| best_oa_location.source.issn | |
| best_oa_location.source.type | repository |
| best_oa_location.source.is_oa | True |
| best_oa_location.source.issn_l | |
| best_oa_location.source.is_core | False |
| best_oa_location.source.is_in_doaj | False |
| best_oa_location.source.display_name | arXiv (Cornell University) |
| best_oa_location.source.host_organization | https://openalex.org/I205783295 |
| best_oa_location.source.host_organization_name | Cornell University |
| best_oa_location.source.host_organization_lineage | https://openalex.org/I205783295 |
| best_oa_location.license | |
| best_oa_location.pdf_url | https://arxiv.org/pdf/2509.22876 |
| best_oa_location.version | submittedVersion |
| best_oa_location.raw_type | text |
| best_oa_location.license_id | |
| best_oa_location.is_accepted | False |
| best_oa_location.is_published | False |
| best_oa_location.raw_source_name | |
| best_oa_location.landing_page_url | http://arxiv.org/abs/2509.22876 |
| primary_location.id | pmh:oai:arXiv.org:2509.22876 |
| primary_location.is_oa | True |
| primary_location.source.id | https://openalex.org/S4306400194 |
| primary_location.source.issn | |
| primary_location.source.type | repository |
| primary_location.source.is_oa | True |
| primary_location.source.issn_l | |
| primary_location.source.is_core | False |
| primary_location.source.is_in_doaj | False |
| primary_location.source.display_name | arXiv (Cornell University) |
| primary_location.source.host_organization | https://openalex.org/I205783295 |
| primary_location.source.host_organization_name | Cornell University |
| primary_location.source.host_organization_lineage | https://openalex.org/I205783295 |
| primary_location.license | |
| primary_location.pdf_url | https://arxiv.org/pdf/2509.22876 |
| primary_location.version | submittedVersion |
| primary_location.raw_type | text |
| primary_location.license_id | |
| primary_location.is_accepted | False |
| primary_location.is_published | False |
| primary_location.raw_source_name | |
| primary_location.landing_page_url | http://arxiv.org/abs/2509.22876 |
| publication_date | 2025-09-26 |
| publication_year | 2025 |
| referenced_works_count | 0 |
| abstract_inverted_index.a | 70, 75, 138, 175, 182, 194 |
| abstract_inverted_index.By | 94 |
| abstract_inverted_index.In | 181 |
| abstract_inverted_index.We | 120 |
| abstract_inverted_index.an | 145 |
| abstract_inverted_index.as | 24, 193 |
| abstract_inverted_index.by | 44, 90, 144 |
| abstract_inverted_index.do | 34 |
| abstract_inverted_index.in | 6, 162, 207, 214, 219 |
| abstract_inverted_index.it | 185 |
| abstract_inverted_index.of | 10, 40, 78, 100, 225 |
| abstract_inverted_index.on | 13, 28, 69, 84, 124 |
| abstract_inverted_index.or | 30 |
| abstract_inverted_index.to | 110, 157, 187 |
| abstract_inverted_index.we | 54, 172 |
| abstract_inverted_index.Dr. | 91 |
| abstract_inverted_index.Our | 135, 200 |
| abstract_inverted_index.and | 115, 133, 159, 221 |
| abstract_inverted_index.but | 217 |
| abstract_inverted_index.can | 50 |
| abstract_inverted_index.for | 63, 178, 197 |
| abstract_inverted_index.has | 2 |
| abstract_inverted_index.key | 195 |
| abstract_inverted_index.lie | 211 |
| abstract_inverted_index.may | 210 |
| abstract_inverted_index.new | 140 |
| abstract_inverted_index.not | 35, 212 |
| abstract_inverted_index.our | 105, 122 |
| abstract_inverted_index.set | 77 |
| abstract_inverted_index.six | 86 |
| abstract_inverted_index.the | 8, 37, 85, 97, 101, 108, 168, 204, 223, 226 |
| abstract_inverted_index.Last | 131 |
| abstract_inverted_index.Paul | 92 |
| abstract_inverted_index.They | 33 |
| abstract_inverted_index.also | 173, 218 |
| abstract_inverted_index.just | 213 |
| abstract_inverted_index.more | 117 |
| abstract_inverted_index.next | 205 |
| abstract_inverted_index.over | 164 |
| abstract_inverted_index.same | 169 |
| abstract_inverted_index.such | 23 |
| abstract_inverted_index.that | 48, 59, 203 |
| abstract_inverted_index.this | 148 |
| abstract_inverted_index.tone | 99 |
| abstract_inverted_index.uses | 60 |
| abstract_inverted_index.when | 142 |
| abstract_inverted_index.with | 167 |
| abstract_inverted_index.Exam, | 132 |
| abstract_inverted_index.HEART | 66 |
| abstract_inverted_index.based | 83 |
| abstract_inverted_index.focus | 27 |
| abstract_inverted_index.gains | 190 |
| abstract_inverted_index.model | 109 |
| abstract_inverted_index.novel | 57 |
| abstract_inverted_index.paths | 114 |
| abstract_inverted_index.shown | 3 |
| abstract_inverted_index.tasks | 16 |
| abstract_inverted_index.these | 189 |
| abstract_inverted_index.using | 74 |
| abstract_inverted_index.work. | 199 |
| abstract_inverted_index.Ekman. | 93 |
| abstract_inverted_index.across | 103 |
| abstract_inverted_index.deeper | 154 |
| abstract_inverted_index.escape | 111 |
| abstract_inverted_index.flawed | 112 |
| abstract_inverted_index.future | 198 |
| abstract_inverted_index.guided | 143 |
| abstract_inverted_index.guides | 107 |
| abstract_inverted_index.logic, | 216 |
| abstract_inverted_index.method | 106 |
| abstract_inverted_index.models | 12 |
| abstract_inverted_index.oracle | 146 |
| abstract_inverted_index.reveal | 137 |
| abstract_inverted_index.`HEART' | 224 |
| abstract_inverted_index.charged | 81 |
| abstract_inverted_index.complex | 14 |
| abstract_inverted_index.curated | 76 |
| abstract_inverted_index.current | 21 |
| abstract_inverted_index.explore | 116 |
| abstract_inverted_index.guiding | 38 |
| abstract_inverted_index.harness | 188 |
| abstract_inverted_index.leading | 156 |
| abstract_inverted_index.logical | 29 |
| abstract_inverted_index.machine | 208 |
| abstract_inverted_index.model's | 71 |
| abstract_inverted_index.models. | 227 |
| abstract_inverted_index.phrases | 82 |
| abstract_inverted_index.prompts | 62 |
| abstract_inverted_index.results | 136 |
| abstract_inverted_index.scaling | 1 |
| abstract_inverted_index.showing | 47 |
| abstract_inverted_index.success | 5 |
| abstract_inverted_index.suggest | 202 |
| abstract_inverted_index.unlocks | 152 |
| abstract_inverted_index.varying | 96 |
| abstract_inverted_index.without | 17 |
| abstract_inverted_index.HEART--a | 56 |
| abstract_inverted_index.However, | 20, 171 |
| abstract_inverted_index.Inspired | 43 |
| abstract_inverted_index.accuracy | 163 |
| abstract_inverted_index.concise, | 79 |
| abstract_inverted_index.critical | 176 |
| abstract_inverted_index.emotions | 49, 88 |
| abstract_inverted_index.evaluate | 121 |
| abstract_inverted_index.feedback | 68, 102 |
| abstract_inverted_index.findings | 201 |
| abstract_inverted_index.frontier | 206 |
| abstract_inverted_index.identify | 174 |
| abstract_inverted_index.language | 11 |
| abstract_inverted_index.leverage | 36 |
| abstract_inverted_index.modulate | 51 |
| abstract_inverted_index.protocol | 151 |
| abstract_inverted_index.provides | 67 |
| abstract_inverted_index.refining | 215 |
| abstract_inverted_index.research | 46 |
| abstract_inverted_index.response | 73 |
| abstract_inverted_index.setting, | 184 |
| abstract_inverted_index.SimpleQA. | 134 |
| abstract_inverted_index.Test-time | 0 |
| abstract_inverted_index.affective | 41, 149 |
| abstract_inverted_index.baselines | 166 |
| abstract_inverted_index.challenge | 196 |
| abstract_inverted_index.cognitive | 52 |
| abstract_inverted_index.emotional | 98 |
| abstract_inverted_index.feedback. | 42 |
| abstract_inverted_index.framework | 58, 123 |
| abstract_inverted_index.improving | 7 |
| abstract_inverted_index.including | 128 |
| abstract_inverted_index.incorrect | 72 |
| abstract_inverted_index.increases | 161 |
| abstract_inverted_index.introduce | 55 |
| abstract_inverted_index.iteration | 150 |
| abstract_inverted_index.iterative | 64 |
| abstract_inverted_index.potential | 39 |
| abstract_inverted_index.practical | 179 |
| abstract_inverted_index.primarily | 26 |
| abstract_inverted_index.promising | 118 |
| abstract_inverted_index.reasoning | 15, 113, 126, 209 |
| abstract_inverted_index.requiring | 18 |
| abstract_inverted_index.struggles | 186 |
| abstract_inverted_index.universal | 87 |
| abstract_inverted_index.verifier, | 147 |
| abstract_inverted_index.verifier. | 170 |
| abstract_inverted_index.Humanity's | 130 |
| abstract_inverted_index.benchmarks | 127 |
| abstract_inverted_index.bottleneck | 177 |
| abstract_inverted_index.consistent | 158 |
| abstract_inverted_index.leveraging | 222 |
| abstract_inverted_index.reasoning, | 155 |
| abstract_inverted_index.strategies | 22 |
| abstract_inverted_index.structural | 31 |
| abstract_inverted_index.categorized | 89 |
| abstract_inverted_index.challenging | 125 |
| abstract_inverted_index.deployment. | 180 |
| abstract_inverted_index.emotionally | 80 |
| abstract_inverted_index.iterations, | 104 |
| abstract_inverted_index.performance | 9 |
| abstract_inverted_index.phenomenon: | 141 |
| abstract_inverted_index.refinement. | 32 |
| abstract_inverted_index.significant | 139 |
| abstract_inverted_index.substantial | 160 |
| abstract_inverted_index.considerable | 4 |
| abstract_inverted_index.fine-tuning. | 19 |
| abstract_inverted_index.highlighting | 192 |
| abstract_inverted_index.performance, | 53 |
| abstract_inverted_index.alternatives. | 119 |
| abstract_inverted_index.consistently, | 191 |
| abstract_inverted_index.psychological | 45 |
| abstract_inverted_index.significantly | 153 |
| abstract_inverted_index.understanding | 220 |
| abstract_inverted_index.verifier-free | 183 |
| abstract_inverted_index.OlympiadBench, | 129 |
| abstract_inverted_index.systematically | 95 |
| abstract_inverted_index.self-reflection | 25 |
| abstract_inverted_index.self-correction. | 65 |
| abstract_inverted_index.state-of-the-art | 165 |
| abstract_inverted_index.emotionally-driven | 61 |
| cited_by_percentile_year | |
| countries_distinct_count | 0 |
| institutions_distinct_count | 7 |
| citation_normalized_percentile |