Fairer Preferences Elicit Improved Human-Aligned Large Language Model Judgments Article Swipe
Han Zhou
,
Xingchen Wan
,
Yinhong Liu
,
Nigel Collier
,
Ivan Vulić
,
Anna Korhonen
·
YOU?
·
· 2024
· Open Access
·
· DOI: https://doi.org/10.48550/arxiv.2406.11370
YOU?
·
· 2024
· Open Access
·
· DOI: https://doi.org/10.48550/arxiv.2406.11370
Large language models (LLMs) have shown promising abilities as cost-effective and reference-free evaluators for assessing language generation quality. In particular, pairwise LLM evaluators, which compare two generated texts and determine the preferred one, have been employed in a wide range of applications. However, LLMs exhibit preference biases and worrying sensitivity to prompt designs. In this work, we first reveal that the predictive preference of LLMs can be highly brittle and skewed, even with semantically equivalent instructions. We find that fairer predictive preferences from LLMs consistently lead to judgments that are better aligned with humans. Motivated by this phenomenon, we propose an automatic Zero-shot Evaluation-oriented Prompt Optimization framework, ZEPO, which aims to produce fairer preference decisions and improve the alignment of LLM evaluators with human judgments. To this end, we propose a zero-shot learning objective based on the preference decision fairness. ZEPO demonstrates substantial performance improvements over state-of-the-art LLM evaluators, without requiring labeled data, on representative meta-evaluation benchmarks. Our findings underscore the critical correlation between preference fairness and human alignment, positioning ZEPO as an efficient prompt optimizer for bridging the gap between LLM evaluators and human judgments.
Related Topics
Metadata
- Type
- preprint
- Language
- en
- Landing Page
- http://arxiv.org/abs/2406.11370
- https://arxiv.org/pdf/2406.11370
- OA Status
- green
- Related Works
- 10
- OpenAlex ID
- https://openalex.org/W4399795120
All OpenAlex metadata
Raw OpenAlex JSON
- OpenAlex ID
-
https://openalex.org/W4399795120Canonical identifier for this work in OpenAlex
- DOI
-
https://doi.org/10.48550/arxiv.2406.11370Digital Object Identifier
- Title
-
Fairer Preferences Elicit Improved Human-Aligned Large Language Model JudgmentsWork title
- Type
-
preprintOpenAlex work type
- Language
-
enPrimary language
- Publication year
-
2024Year of publication
- Publication date
-
2024-06-17Full publication date if available
- Authors
-
Han Zhou, Xingchen Wan, Yinhong Liu, Nigel Collier, Ivan Vulić, Anna KorhonenList of authors in order
- Landing page
-
https://arxiv.org/abs/2406.11370Publisher landing page
- PDF URL
-
https://arxiv.org/pdf/2406.11370Direct 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/2406.11370Direct OA link when available
- Concepts
-
Psychology, Business, Computer science, Cognitive psychology, Social psychologyTop concepts (fields/topics) attached by OpenAlex
- Cited by
-
0Total citation count in OpenAlex
- Related works (count)
-
10Other works algorithmically related by OpenAlex
Full payload
| id | https://openalex.org/W4399795120 |
|---|---|
| doi | https://doi.org/10.48550/arxiv.2406.11370 |
| ids.doi | https://doi.org/10.48550/arxiv.2406.11370 |
| ids.openalex | https://openalex.org/W4399795120 |
| fwci | 0.0 |
| type | preprint |
| title | Fairer Preferences Elicit Improved Human-Aligned Large Language Model Judgments |
| 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 | 0.84579998254776 |
| 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 | 0.7800999879837036 |
| 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 |
| topics[2].id | https://openalex.org/T12031 |
| topics[2].field.id | https://openalex.org/fields/17 |
| topics[2].field.display_name | Computer Science |
| topics[2].score | 0.7113000154495239 |
| topics[2].domain.id | https://openalex.org/domains/3 |
| topics[2].domain.display_name | Physical Sciences |
| topics[2].subfield.id | https://openalex.org/subfields/1702 |
| topics[2].subfield.display_name | Artificial Intelligence |
| topics[2].display_name | Speech and dialogue systems |
| is_xpac | False |
| apc_list | |
| apc_paid | |
| concepts[0].id | https://openalex.org/C15744967 |
| concepts[0].level | 0 |
| concepts[0].score | 0.4500669240951538 |
| concepts[0].wikidata | https://www.wikidata.org/wiki/Q9418 |
| concepts[0].display_name | Psychology |
| concepts[1].id | https://openalex.org/C144133560 |
| concepts[1].level | 0 |
| concepts[1].score | 0.4097387194633484 |
| concepts[1].wikidata | https://www.wikidata.org/wiki/Q4830453 |
| concepts[1].display_name | Business |
| concepts[2].id | https://openalex.org/C41008148 |
| concepts[2].level | 0 |
| concepts[2].score | 0.39968395233154297 |
| concepts[2].wikidata | https://www.wikidata.org/wiki/Q21198 |
| concepts[2].display_name | Computer science |
| concepts[3].id | https://openalex.org/C180747234 |
| concepts[3].level | 1 |
| concepts[3].score | 0.3404689431190491 |
| concepts[3].wikidata | https://www.wikidata.org/wiki/Q23373 |
| concepts[3].display_name | Cognitive psychology |
| concepts[4].id | https://openalex.org/C77805123 |
| concepts[4].level | 1 |
| concepts[4].score | 0.33380791544914246 |
| concepts[4].wikidata | https://www.wikidata.org/wiki/Q161272 |
| concepts[4].display_name | Social psychology |
| keywords[0].id | https://openalex.org/keywords/psychology |
| keywords[0].score | 0.4500669240951538 |
| keywords[0].display_name | Psychology |
| keywords[1].id | https://openalex.org/keywords/business |
| keywords[1].score | 0.4097387194633484 |
| keywords[1].display_name | Business |
| keywords[2].id | https://openalex.org/keywords/computer-science |
| keywords[2].score | 0.39968395233154297 |
| keywords[2].display_name | Computer science |
| keywords[3].id | https://openalex.org/keywords/cognitive-psychology |
| keywords[3].score | 0.3404689431190491 |
| keywords[3].display_name | Cognitive psychology |
| keywords[4].id | https://openalex.org/keywords/social-psychology |
| keywords[4].score | 0.33380791544914246 |
| keywords[4].display_name | Social psychology |
| language | en |
| locations[0].id | pmh:oai:arXiv.org:2406.11370 |
| 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/2406.11370 |
| 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/2406.11370 |
| locations[1].id | doi:10.48550/arxiv.2406.11370 |
| 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 | |
| locations[1].pdf_url | |
| locations[1].version | |
| locations[1].raw_type | article |
| locations[1].license_id | |
| 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.2406.11370 |
| indexed_in | arxiv, datacite |
| authorships[0].author.id | https://openalex.org/A5100784571 |
| authorships[0].author.orcid | https://orcid.org/0000-0003-3778-4075 |
| authorships[0].author.display_name | Han Zhou |
| authorships[0].author_position | first |
| authorships[0].raw_author_name | Zhou, Han |
| authorships[0].is_corresponding | False |
| authorships[1].author.id | https://openalex.org/A5083203503 |
| authorships[1].author.orcid | https://orcid.org/0000-0003-0074-0597 |
| authorships[1].author.display_name | Xingchen Wan |
| authorships[1].author_position | middle |
| authorships[1].raw_author_name | Wan, Xingchen |
| authorships[1].is_corresponding | False |
| authorships[2].author.id | https://openalex.org/A5014118666 |
| authorships[2].author.orcid | https://orcid.org/0009-0006-6723-0558 |
| authorships[2].author.display_name | Yinhong Liu |
| authorships[2].author_position | middle |
| authorships[2].raw_author_name | Liu, Yinhong |
| authorships[2].is_corresponding | False |
| authorships[3].author.id | https://openalex.org/A5073413742 |
| authorships[3].author.orcid | https://orcid.org/0000-0002-7230-4164 |
| authorships[3].author.display_name | Nigel Collier |
| authorships[3].author_position | middle |
| authorships[3].raw_author_name | Collier, Nigel |
| authorships[3].is_corresponding | False |
| authorships[4].author.id | https://openalex.org/A5014866912 |
| authorships[4].author.orcid | https://orcid.org/0000-0002-5161-5422 |
| authorships[4].author.display_name | Ivan Vulić |
| authorships[4].author_position | middle |
| authorships[4].raw_author_name | Vulić, Ivan |
| authorships[4].is_corresponding | False |
| authorships[5].author.id | https://openalex.org/A5081393566 |
| authorships[5].author.orcid | https://orcid.org/0000-0002-3692-3144 |
| authorships[5].author.display_name | Anna Korhonen |
| authorships[5].author_position | last |
| authorships[5].raw_author_name | Korhonen, Anna |
| authorships[5].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/2406.11370 |
| open_access.oa_status | green |
| open_access.any_repository_has_fulltext | False |
| created_date | 2025-10-10T00:00:00 |
| display_name | Fairer Preferences Elicit Improved Human-Aligned Large Language Model Judgments |
| has_fulltext | False |
| 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 | 0.84579998254776 |
| 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 |
| related_works | https://openalex.org/W2748952813, https://openalex.org/W4391375266, https://openalex.org/W2390279801, https://openalex.org/W2358668433, https://openalex.org/W4396701345, https://openalex.org/W2376932109, https://openalex.org/W2001405890, https://openalex.org/W4396696052, https://openalex.org/W2382290278, https://openalex.org/W2010073985 |
| cited_by_count | 0 |
| locations_count | 2 |
| best_oa_location.id | pmh:oai:arXiv.org:2406.11370 |
| 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/2406.11370 |
| 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/2406.11370 |
| primary_location.id | pmh:oai:arXiv.org:2406.11370 |
| 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/2406.11370 |
| 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/2406.11370 |
| publication_date | 2024-06-17 |
| publication_year | 2024 |
| referenced_works_count | 0 |
| abstract_inverted_index.a | 37, 130 |
| abstract_inverted_index.In | 18, 53 |
| abstract_inverted_index.To | 125 |
| abstract_inverted_index.We | 76 |
| abstract_inverted_index.an | 100, 172 |
| abstract_inverted_index.as | 8, 171 |
| abstract_inverted_index.be | 66 |
| abstract_inverted_index.by | 95 |
| abstract_inverted_index.in | 36 |
| abstract_inverted_index.of | 40, 63, 119 |
| abstract_inverted_index.on | 135, 153 |
| abstract_inverted_index.to | 50, 86, 110 |
| abstract_inverted_index.we | 56, 98, 128 |
| abstract_inverted_index.LLM | 21, 120, 147, 181 |
| abstract_inverted_index.Our | 157 |
| abstract_inverted_index.and | 10, 28, 47, 69, 115, 166, 183 |
| abstract_inverted_index.are | 89 |
| abstract_inverted_index.can | 65 |
| abstract_inverted_index.for | 13, 176 |
| abstract_inverted_index.gap | 179 |
| abstract_inverted_index.the | 30, 60, 117, 136, 160, 178 |
| abstract_inverted_index.two | 25 |
| abstract_inverted_index.LLMs | 43, 64, 83 |
| abstract_inverted_index.ZEPO | 140, 170 |
| abstract_inverted_index.aims | 109 |
| abstract_inverted_index.been | 34 |
| abstract_inverted_index.end, | 127 |
| abstract_inverted_index.even | 71 |
| abstract_inverted_index.find | 77 |
| abstract_inverted_index.from | 82 |
| abstract_inverted_index.have | 4, 33 |
| abstract_inverted_index.lead | 85 |
| abstract_inverted_index.one, | 32 |
| abstract_inverted_index.over | 145 |
| abstract_inverted_index.that | 59, 78, 88 |
| abstract_inverted_index.this | 54, 96, 126 |
| abstract_inverted_index.wide | 38 |
| abstract_inverted_index.with | 72, 92, 122 |
| abstract_inverted_index.Large | 0 |
| abstract_inverted_index.ZEPO, | 107 |
| abstract_inverted_index.based | 134 |
| abstract_inverted_index.data, | 152 |
| abstract_inverted_index.first | 57 |
| abstract_inverted_index.human | 123, 167, 184 |
| abstract_inverted_index.range | 39 |
| abstract_inverted_index.shown | 5 |
| abstract_inverted_index.texts | 27 |
| abstract_inverted_index.which | 23, 108 |
| abstract_inverted_index.work, | 55 |
| abstract_inverted_index.(LLMs) | 3 |
| abstract_inverted_index.Prompt | 104 |
| abstract_inverted_index.better | 90 |
| abstract_inverted_index.biases | 46 |
| abstract_inverted_index.fairer | 79, 112 |
| abstract_inverted_index.highly | 67 |
| abstract_inverted_index.models | 2 |
| abstract_inverted_index.prompt | 51, 174 |
| abstract_inverted_index.reveal | 58 |
| abstract_inverted_index.aligned | 91 |
| abstract_inverted_index.between | 163, 180 |
| abstract_inverted_index.brittle | 68 |
| abstract_inverted_index.compare | 24 |
| abstract_inverted_index.exhibit | 44 |
| abstract_inverted_index.humans. | 93 |
| abstract_inverted_index.improve | 116 |
| abstract_inverted_index.labeled | 151 |
| abstract_inverted_index.produce | 111 |
| abstract_inverted_index.propose | 99, 129 |
| abstract_inverted_index.skewed, | 70 |
| abstract_inverted_index.without | 149 |
| abstract_inverted_index.However, | 42 |
| abstract_inverted_index.bridging | 177 |
| abstract_inverted_index.critical | 161 |
| abstract_inverted_index.decision | 138 |
| abstract_inverted_index.designs. | 52 |
| abstract_inverted_index.employed | 35 |
| abstract_inverted_index.fairness | 165 |
| abstract_inverted_index.findings | 158 |
| abstract_inverted_index.language | 1, 15 |
| abstract_inverted_index.learning | 132 |
| abstract_inverted_index.pairwise | 20 |
| abstract_inverted_index.quality. | 17 |
| abstract_inverted_index.worrying | 48 |
| abstract_inverted_index.Motivated | 94 |
| abstract_inverted_index.Zero-shot | 102 |
| abstract_inverted_index.abilities | 7 |
| abstract_inverted_index.alignment | 118 |
| abstract_inverted_index.assessing | 14 |
| abstract_inverted_index.automatic | 101 |
| abstract_inverted_index.decisions | 114 |
| abstract_inverted_index.determine | 29 |
| abstract_inverted_index.efficient | 173 |
| abstract_inverted_index.fairness. | 139 |
| abstract_inverted_index.generated | 26 |
| abstract_inverted_index.judgments | 87 |
| abstract_inverted_index.objective | 133 |
| abstract_inverted_index.optimizer | 175 |
| abstract_inverted_index.preferred | 31 |
| abstract_inverted_index.promising | 6 |
| abstract_inverted_index.requiring | 150 |
| abstract_inverted_index.zero-shot | 131 |
| abstract_inverted_index.alignment, | 168 |
| abstract_inverted_index.equivalent | 74 |
| abstract_inverted_index.evaluators | 12, 121, 182 |
| abstract_inverted_index.framework, | 106 |
| abstract_inverted_index.generation | 16 |
| abstract_inverted_index.judgments. | 124, 185 |
| abstract_inverted_index.predictive | 61, 80 |
| abstract_inverted_index.preference | 45, 62, 113, 137, 164 |
| abstract_inverted_index.underscore | 159 |
| abstract_inverted_index.benchmarks. | 156 |
| abstract_inverted_index.correlation | 162 |
| abstract_inverted_index.evaluators, | 22, 148 |
| abstract_inverted_index.particular, | 19 |
| abstract_inverted_index.performance | 143 |
| abstract_inverted_index.phenomenon, | 97 |
| abstract_inverted_index.positioning | 169 |
| abstract_inverted_index.preferences | 81 |
| abstract_inverted_index.sensitivity | 49 |
| abstract_inverted_index.substantial | 142 |
| abstract_inverted_index.Optimization | 105 |
| abstract_inverted_index.consistently | 84 |
| abstract_inverted_index.demonstrates | 141 |
| abstract_inverted_index.improvements | 144 |
| abstract_inverted_index.semantically | 73 |
| abstract_inverted_index.applications. | 41 |
| abstract_inverted_index.instructions. | 75 |
| abstract_inverted_index.cost-effective | 9 |
| abstract_inverted_index.reference-free | 11 |
| abstract_inverted_index.representative | 154 |
| abstract_inverted_index.meta-evaluation | 155 |
| abstract_inverted_index.state-of-the-art | 146 |
| abstract_inverted_index.Evaluation-oriented | 103 |
| cited_by_percentile_year | |
| countries_distinct_count | 0 |
| institutions_distinct_count | 6 |
| citation_normalized_percentile |