Guiding Reasoning in Small Language Models with LLM Assistance Article Swipe
YOU?
·
· 2025
· Open Access
·
· DOI: https://doi.org/10.48550/arxiv.2504.09923
The limited reasoning capabilities of small language models (SLMs) cast doubt on their suitability for tasks demanding deep, multi-step logical deduction. This paper introduces a framework called Small Reasons, Large Hints (SMART), which selectively augments SLM reasoning with targeted guidance from large language models (LLMs). Inspired by the concept of cognitive scaffolding, SMART employs a score-based evaluation to identify uncertain reasoning steps and injects corrective LLM-generated reasoning only when necessary. By framing structured reasoning as an optimal policy search, our approach steers the reasoning trajectory toward correct solutions without exhaustive sampling. Our experiments on mathematical reasoning datasets demonstrate that targeted external scaffolding significantly improves performance, paving the way for collaborative use of both SLM and LLM to tackle complex reasoning tasks that are currently unsolvable by SLMs alone.
Related Topics
- Type
- preprint
- Language
- en
- Landing Page
- http://arxiv.org/abs/2504.09923
- https://arxiv.org/pdf/2504.09923
- OA Status
- green
- OpenAlex ID
- https://openalex.org/W4415158827
Raw OpenAlex JSON
- OpenAlex ID
-
https://openalex.org/W4415158827Canonical identifier for this work in OpenAlex
- DOI
-
https://doi.org/10.48550/arxiv.2504.09923Digital Object Identifier
- Title
-
Guiding Reasoning in Small Language Models with LLM AssistanceWork title
- Type
-
preprintOpenAlex work type
- Language
-
enPrimary language
- Publication year
-
2025Year of publication
- Publication date
-
2025-04-14Full publication date if available
- Authors
-
Yujin Kim, Euiin Yi, M. Kim, Se-Young Yun, Taehyeon KimList of authors in order
- Landing page
-
https://arxiv.org/abs/2504.09923Publisher landing page
- PDF URL
-
https://arxiv.org/pdf/2504.09923Direct 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/2504.09923Direct OA link when available
- Cited by
-
0Total citation count in OpenAlex
Full payload
| id | https://openalex.org/W4415158827 |
|---|---|
| doi | https://doi.org/10.48550/arxiv.2504.09923 |
| ids.doi | https://doi.org/10.48550/arxiv.2504.09923 |
| ids.openalex | https://openalex.org/W4415158827 |
| fwci | |
| type | preprint |
| title | Guiding Reasoning in Small Language Models with LLM Assistance |
| biblio.issue | |
| biblio.volume | |
| biblio.last_page | |
| biblio.first_page | |
| topics[0].id | https://openalex.org/T10181 |
| topics[0].field.id | https://openalex.org/fields/17 |
| topics[0].field.display_name | Computer Science |
| topics[0].score | 0.9947999715805054 |
| 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 | Natural Language Processing Techniques |
| topics[1].id | https://openalex.org/T10028 |
| topics[1].field.id | https://openalex.org/fields/17 |
| topics[1].field.display_name | Computer Science |
| topics[1].score | 0.9611999988555908 |
| topics[1].domain.id | https://openalex.org/domains/3 |
| topics[1].domain.display_name | Physical Sciences |
| topics[1].subfield.id | https://openalex.org/subfields/1702 |
| topics[1].subfield.display_name | Artificial Intelligence |
| topics[1].display_name | Topic Modeling |
| is_xpac | False |
| apc_list | |
| apc_paid | |
| language | en |
| locations[0].id | pmh:oai:arXiv.org:2504.09923 |
| 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 | cc-by |
| locations[0].pdf_url | https://arxiv.org/pdf/2504.09923 |
| locations[0].version | submittedVersion |
| locations[0].raw_type | text |
| locations[0].license_id | https://openalex.org/licenses/cc-by |
| locations[0].is_accepted | False |
| locations[0].is_published | False |
| locations[0].raw_source_name | |
| locations[0].landing_page_url | http://arxiv.org/abs/2504.09923 |
| locations[1].id | doi:10.48550/arxiv.2504.09923 |
| 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.2504.09923 |
| indexed_in | arxiv, datacite |
| authorships[0].author.id | https://openalex.org/A5100461241 |
| authorships[0].author.orcid | https://orcid.org/0000-0003-4261-1793 |
| authorships[0].author.display_name | Yujin Kim |
| authorships[0].author_position | first |
| authorships[0].raw_author_name | Kim, Yujin |
| authorships[0].is_corresponding | False |
| authorships[1].author.id | https://openalex.org/A5119990043 |
| authorships[1].author.orcid | |
| authorships[1].author.display_name | Euiin Yi |
| authorships[1].author_position | middle |
| authorships[1].raw_author_name | Yi, Euiin |
| authorships[1].is_corresponding | False |
| authorships[2].author.id | https://openalex.org/A5041821271 |
| authorships[2].author.orcid | https://orcid.org/0000-0002-0906-062X |
| authorships[2].author.display_name | M. Kim |
| authorships[2].author_position | middle |
| authorships[2].raw_author_name | Kim, Minu |
| authorships[2].is_corresponding | False |
| authorships[3].author.id | https://openalex.org/A5091674853 |
| authorships[3].author.orcid | |
| authorships[3].author.display_name | Se-Young Yun |
| authorships[3].author_position | middle |
| authorships[3].raw_author_name | Yun, Se-Young |
| authorships[3].is_corresponding | False |
| authorships[4].author.id | https://openalex.org/A5100774197 |
| authorships[4].author.orcid | https://orcid.org/0000-0001-5496-2625 |
| authorships[4].author.display_name | Taehyeon Kim |
| authorships[4].author_position | last |
| authorships[4].raw_author_name | Kim, Taehyeon |
| authorships[4].is_corresponding | False |
| has_content.pdf | True |
| has_content.grobid_xml | True |
| is_paratext | False |
| open_access.is_oa | True |
| open_access.oa_url | https://arxiv.org/pdf/2504.09923 |
| open_access.oa_status | green |
| open_access.any_repository_has_fulltext | False |
| created_date | 2025-10-14T00:00:00 |
| display_name | Guiding Reasoning in Small Language Models with LLM Assistance |
| has_fulltext | True |
| is_retracted | False |
| updated_date | 2025-11-06T06:51:31.235846 |
| primary_topic.id | https://openalex.org/T10181 |
| primary_topic.field.id | https://openalex.org/fields/17 |
| primary_topic.field.display_name | Computer Science |
| primary_topic.score | 0.9947999715805054 |
| 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 | Natural Language Processing Techniques |
| cited_by_count | 0 |
| locations_count | 2 |
| best_oa_location.id | pmh:oai:arXiv.org:2504.09923 |
| 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 | cc-by |
| best_oa_location.pdf_url | https://arxiv.org/pdf/2504.09923 |
| best_oa_location.version | submittedVersion |
| best_oa_location.raw_type | text |
| best_oa_location.license_id | https://openalex.org/licenses/cc-by |
| 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/2504.09923 |
| primary_location.id | pmh:oai:arXiv.org:2504.09923 |
| 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 | cc-by |
| primary_location.pdf_url | https://arxiv.org/pdf/2504.09923 |
| primary_location.version | submittedVersion |
| primary_location.raw_type | text |
| primary_location.license_id | https://openalex.org/licenses/cc-by |
| primary_location.is_accepted | False |
| primary_location.is_published | False |
| primary_location.raw_source_name | |
| primary_location.landing_page_url | http://arxiv.org/abs/2504.09923 |
| publication_date | 2025-04-14 |
| publication_year | 2025 |
| referenced_works_count | 0 |
| abstract_inverted_index.a | 24, 54 |
| abstract_inverted_index.By | 70 |
| abstract_inverted_index.an | 75 |
| abstract_inverted_index.as | 74 |
| abstract_inverted_index.by | 46, 125 |
| abstract_inverted_index.of | 4, 49, 111 |
| abstract_inverted_index.on | 11, 93 |
| abstract_inverted_index.to | 57, 116 |
| abstract_inverted_index.LLM | 115 |
| abstract_inverted_index.Our | 91 |
| abstract_inverted_index.SLM | 35, 113 |
| abstract_inverted_index.The | 0 |
| abstract_inverted_index.and | 62, 114 |
| abstract_inverted_index.are | 122 |
| abstract_inverted_index.for | 14, 108 |
| abstract_inverted_index.our | 79 |
| abstract_inverted_index.the | 47, 82, 106 |
| abstract_inverted_index.use | 110 |
| abstract_inverted_index.way | 107 |
| abstract_inverted_index.SLMs | 126 |
| abstract_inverted_index.This | 21 |
| abstract_inverted_index.both | 112 |
| abstract_inverted_index.cast | 9 |
| abstract_inverted_index.from | 40 |
| abstract_inverted_index.only | 67 |
| abstract_inverted_index.that | 98, 121 |
| abstract_inverted_index.when | 68 |
| abstract_inverted_index.with | 37 |
| abstract_inverted_index.Hints | 30 |
| abstract_inverted_index.Large | 29 |
| abstract_inverted_index.SMART | 52 |
| abstract_inverted_index.Small | 27 |
| abstract_inverted_index.deep, | 17 |
| abstract_inverted_index.doubt | 10 |
| abstract_inverted_index.large | 41 |
| abstract_inverted_index.paper | 22 |
| abstract_inverted_index.small | 5 |
| abstract_inverted_index.steps | 61 |
| abstract_inverted_index.tasks | 15, 120 |
| abstract_inverted_index.their | 12 |
| abstract_inverted_index.which | 32 |
| abstract_inverted_index.(SLMs) | 8 |
| abstract_inverted_index.alone. | 127 |
| abstract_inverted_index.called | 26 |
| abstract_inverted_index.models | 7, 43 |
| abstract_inverted_index.paving | 105 |
| abstract_inverted_index.policy | 77 |
| abstract_inverted_index.steers | 81 |
| abstract_inverted_index.tackle | 117 |
| abstract_inverted_index.toward | 85 |
| abstract_inverted_index.(LLMs). | 44 |
| abstract_inverted_index.complex | 118 |
| abstract_inverted_index.concept | 48 |
| abstract_inverted_index.correct | 86 |
| abstract_inverted_index.employs | 53 |
| abstract_inverted_index.framing | 71 |
| abstract_inverted_index.injects | 63 |
| abstract_inverted_index.limited | 1 |
| abstract_inverted_index.logical | 19 |
| abstract_inverted_index.optimal | 76 |
| abstract_inverted_index.search, | 78 |
| abstract_inverted_index.without | 88 |
| abstract_inverted_index.(SMART), | 31 |
| abstract_inverted_index.Inspired | 45 |
| abstract_inverted_index.Reasons, | 28 |
| abstract_inverted_index.approach | 80 |
| abstract_inverted_index.augments | 34 |
| abstract_inverted_index.datasets | 96 |
| abstract_inverted_index.external | 100 |
| abstract_inverted_index.guidance | 39 |
| abstract_inverted_index.identify | 58 |
| abstract_inverted_index.improves | 103 |
| abstract_inverted_index.language | 6, 42 |
| abstract_inverted_index.targeted | 38, 99 |
| abstract_inverted_index.cognitive | 50 |
| abstract_inverted_index.currently | 123 |
| abstract_inverted_index.demanding | 16 |
| abstract_inverted_index.framework | 25 |
| abstract_inverted_index.reasoning | 2, 36, 60, 66, 73, 83, 95, 119 |
| abstract_inverted_index.sampling. | 90 |
| abstract_inverted_index.solutions | 87 |
| abstract_inverted_index.uncertain | 59 |
| abstract_inverted_index.corrective | 64 |
| abstract_inverted_index.deduction. | 20 |
| abstract_inverted_index.evaluation | 56 |
| abstract_inverted_index.exhaustive | 89 |
| abstract_inverted_index.introduces | 23 |
| abstract_inverted_index.multi-step | 18 |
| abstract_inverted_index.necessary. | 69 |
| abstract_inverted_index.structured | 72 |
| abstract_inverted_index.trajectory | 84 |
| abstract_inverted_index.unsolvable | 124 |
| abstract_inverted_index.demonstrate | 97 |
| abstract_inverted_index.experiments | 92 |
| abstract_inverted_index.scaffolding | 101 |
| abstract_inverted_index.score-based | 55 |
| abstract_inverted_index.selectively | 33 |
| abstract_inverted_index.suitability | 13 |
| abstract_inverted_index.capabilities | 3 |
| abstract_inverted_index.mathematical | 94 |
| abstract_inverted_index.performance, | 104 |
| abstract_inverted_index.scaffolding, | 51 |
| abstract_inverted_index.LLM-generated | 65 |
| abstract_inverted_index.collaborative | 109 |
| abstract_inverted_index.significantly | 102 |
| cited_by_percentile_year | |
| countries_distinct_count | 0 |
| institutions_distinct_count | 5 |
| citation_normalized_percentile |