Language Model Council: Democratically Benchmarking Foundation Models on Highly Subjective Tasks
As Large Language Models (LLMs) continue to evolve, the search for efficient and meaningful evaluation methods is ongoing. Many recent evaluations use LLMs as judges to score outputs from other LLMs, often relying on a single large model like GPT-4o. However, using a single LLM judge is prone to int...
Saved in:
Published in | arXiv.org |
---|---|
Main Authors | , , , |
Format | Paper |
Language | English |
Published |
Ithaca
Cornell University Library, arXiv.org
21.10.2024
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Abstract | As Large Language Models (LLMs) continue to evolve, the search for efficient and meaningful evaluation methods is ongoing. Many recent evaluations use LLMs as judges to score outputs from other LLMs, often relying on a single large model like GPT-4o. However, using a single LLM judge is prone to intra-model bias, and many tasks - such as those related to emotional intelligence, creative writing, and persuasiveness - may be too subjective for a single model to judge fairly. We introduce the Language Model Council (LMC), where a group of LLMs collaborate to create tests, respond to them, and evaluate each other's responses to produce a ranking in a democratic fashion. Unlike previous approaches that focus on reducing cost or bias by using a panel of smaller models, our work examines the benefits and nuances of a fully inclusive LLM evaluation system. In a detailed case study on emotional intelligence, we deploy a council of 20 recent LLMs to rank each other on open-ended responses to interpersonal conflicts. Our results show that the LMC produces rankings that are more separable and more robust, and through a user study, we show that they are more consistent with human evaluations than any individual LLM judge. Using all LLMs for judging can be costly, however, so we use Monte Carlo simulations and hand-curated sub-councils to study hypothetical council compositions and discuss the value of the incremental LLM judge. |
---|---|
AbstractList | As Large Language Models (LLMs) continue to evolve, the search for efficient and meaningful evaluation methods is ongoing. Many recent evaluations use LLMs as judges to score outputs from other LLMs, often relying on a single large model like GPT-4o. However, using a single LLM judge is prone to intra-model bias, and many tasks - such as those related to emotional intelligence, creative writing, and persuasiveness - may be too subjective for a single model to judge fairly. We introduce the Language Model Council (LMC), where a group of LLMs collaborate to create tests, respond to them, and evaluate each other's responses to produce a ranking in a democratic fashion. Unlike previous approaches that focus on reducing cost or bias by using a panel of smaller models, our work examines the benefits and nuances of a fully inclusive LLM evaluation system. In a detailed case study on emotional intelligence, we deploy a council of 20 recent LLMs to rank each other on open-ended responses to interpersonal conflicts. Our results show that the LMC produces rankings that are more separable and more robust, and through a user study, we show that they are more consistent with human evaluations than any individual LLM judge. Using all LLMs for judging can be costly, however, so we use Monte Carlo simulations and hand-curated sub-councils to study hypothetical council compositions and discuss the value of the incremental LLM judge. |
Author | Plaza-del-Arco, Flor Miriam Genchel, Benjie Zhao, Justin Curry, Amanda Cercas |
Author_xml | – sequence: 1 givenname: Justin surname: Zhao fullname: Zhao, Justin – sequence: 2 givenname: Flor surname: Plaza-del-Arco middlename: Miriam fullname: Plaza-del-Arco, Flor Miriam – sequence: 3 givenname: Benjie surname: Genchel fullname: Genchel, Benjie – sequence: 4 givenname: Amanda surname: Curry middlename: Cercas fullname: Curry, Amanda Cercas |
BookMark | eNqNi9EKgjAYRkcUZOU7DLoWbFOLLrPEi7rKe1vzT2dzK-eC3r5BPUBX34FzvhkaK61ghDxC6SrYRIRMkW9MG4YhSdYkjqmHLkemastqwCddgcSptooLucV76DTv2SA4k_KNd6B407H-LlSNMxdVTmn1fRnsKBd148KzvbbAB_ECXDBzNws0uTFpwP_tHC2zQ5HmwaPXTwtmKFtte-VUScNkQ6JVTAn9r_oAZEJGkA |
ContentType | Paper |
Copyright | 2024. This work is published under http://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License. |
Copyright_xml | – notice: 2024. This work is published under http://creativecommons.org/licenses/by/4.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License. |
DBID | 8FE 8FG ABJCF ABUWG AFKRA AZQEC BENPR BGLVJ CCPQU DWQXO HCIFZ L6V M7S PIMPY PQEST PQQKQ PQUKI PRINS PTHSS |
DatabaseName | ProQuest SciTech Collection ProQuest Technology Collection Materials Science & Engineering Collection ProQuest Central (Alumni) ProQuest Central UK/Ireland ProQuest Central Essentials ProQuest Central Technology Collection ProQuest One Community College ProQuest Central SciTech Premium Collection (Proquest) (PQ_SDU_P3) ProQuest Engineering Collection Engineering Database Publicly Available Content Database ProQuest One Academic Eastern Edition (DO NOT USE) ProQuest One Academic ProQuest One Academic UKI Edition ProQuest Central China Engineering Collection |
DatabaseTitle | Publicly Available Content Database Engineering Database Technology Collection ProQuest Central Essentials ProQuest One Academic Eastern Edition ProQuest Central (Alumni Edition) SciTech Premium Collection ProQuest One Community College ProQuest Technology Collection ProQuest SciTech Collection ProQuest Central China ProQuest Central ProQuest Engineering Collection ProQuest One Academic UKI Edition ProQuest Central Korea Materials Science & Engineering Collection ProQuest One Academic Engineering Collection |
DatabaseTitleList | Publicly Available Content Database |
Database_xml | – sequence: 1 dbid: 8FG name: ProQuest Technology Collection url: https://search.proquest.com/technologycollection1 sourceTypes: Aggregation Database |
DeliveryMethod | fulltext_linktorsrc |
Discipline | Physics |
EISSN | 2331-8422 |
Genre | Working Paper/Pre-Print |
GroupedDBID | 8FE 8FG ABJCF ABUWG AFKRA ALMA_UNASSIGNED_HOLDINGS AZQEC BENPR BGLVJ CCPQU DWQXO FRJ HCIFZ L6V M7S M~E PIMPY PQEST PQQKQ PQUKI PRINS PTHSS |
ID | FETCH-proquest_journals_30682415323 |
IEDL.DBID | BENPR |
IngestDate | Thu Oct 24 04:33:34 EDT 2024 |
IsOpenAccess | true |
IsPeerReviewed | false |
IsScholarly | false |
Language | English |
LinkModel | DirectLink |
MergedId | FETCHMERGED-proquest_journals_30682415323 |
OpenAccessLink | https://www.proquest.com/docview/3068241532?pq-origsite=%requestingapplication% |
PQID | 3068241532 |
PQPubID | 2050157 |
ParticipantIDs | proquest_journals_3068241532 |
PublicationCentury | 2000 |
PublicationDate | 20241021 |
PublicationDateYYYYMMDD | 2024-10-21 |
PublicationDate_xml | – month: 10 year: 2024 text: 20241021 day: 21 |
PublicationDecade | 2020 |
PublicationPlace | Ithaca |
PublicationPlace_xml | – name: Ithaca |
PublicationTitle | arXiv.org |
PublicationYear | 2024 |
Publisher | Cornell University Library, arXiv.org |
Publisher_xml | – name: Cornell University Library, arXiv.org |
SSID | ssj0002672553 |
Score | 3.5695791 |
SecondaryResourceType | preprint |
Snippet | As Large Language Models (LLMs) continue to evolve, the search for efficient and meaningful evaluation methods is ongoing. Many recent evaluations use LLMs as... |
SourceID | proquest |
SourceType | Aggregation Database |
SubjectTerms | Benchmarks Councils Emotional intelligence Large language models Robustness |
Title | Language Model Council: Democratically Benchmarking Foundation Models on Highly Subjective Tasks |
URI | https://www.proquest.com/docview/3068241532 |
hasFullText | 1 |
inHoldings | 1 |
isFullTextHit | |
isPrint | |
link | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwfV1bS8MwFD64FsE3r3iZI6CvwSzpbb4I09YhbgyZsLfZpileunYu88EXf7snXecehL2lpCnNafi-0y8nfACXCHmGhThlQgnquEzQTpa1qccEsmUc-Gll39YfeL1n52HsjmvBTddllStMrIA6LaXRyK8wtQ0M2wh-M_ukxjXK7K7WFhoNsDn-KTAL7G44GD79qSzc8zFnFv-AtmKPaBfsYTxT8z3YUsU-bFdFl1IfwMtjLRYS40iWE3M-XL7l1-ROTUtpPg0GMP8mXVxJr9O4UrXJ2gdpOUoTbJlqDbwRUeB9CWBkFOsPfQgXUTi67dHVa03qpaMn64mKI7CKslDHQIRylMzcII1Z4iQyiTGOCnOfjpdlPkuTE2huetLp5u4z2OHI1QaSebsJ1mL-pc6RaxdJCxpBdN-qw4pX_Z_wF-buiiA |
link.rule.ids | 783,787,12777,21400,33385,33756,43612,43817 |
linkProvider | ProQuest |
linkToHtml | http://utb.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwfV1dT4MwFL3REaNvfsaPqU30tZFRPjZfTKZbUBlZDCZ7w1JKpjKY63zw33vLmHsw2VuTFtJemnMuh0sPwDVCnmYhi5pMMmo7JqOdLGtR12TIlrztpZV92yB0_Vf7aeSMasFN1WWVS0ysgDothdbIbzC1bWu2Ydbd9Itq1yj9dbW20NgEQx9VhS9fRrcXDl_-VBbL9TBnZv-AtmKP_i4YQz6Vsz3YkMU-bFVFl0IdwFtQi4VEO5LlRP8fLt7zW_IgJ6XQjwYDmP-QLu6k8YRXqjZZ-SAtrlIEW7paAwciCnwsAIxEXH2qQ7jq96J7ny6nFddbR8WrhbIjaBRlIY-BMGlLkTntlJuJnYiEYxwl5j4dN8s8M01OoLnuTqfruy9h248GQRw8hs9nsGMhb2t4tlpNaMxn3_IceXeeXNTB_QUU04sD |
openUrl | ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Language+Model+Council%3A+Democratically+Benchmarking+Foundation+Models+on+Highly+Subjective+Tasks&rft.jtitle=arXiv.org&rft.au=Zhao%2C+Justin&rft.au=Plaza-del-Arco%2C+Flor+Miriam&rft.au=Genchel%2C+Benjie&rft.au=Curry%2C+Amanda+Cercas&rft.date=2024-10-21&rft.pub=Cornell+University+Library%2C+arXiv.org&rft.eissn=2331-8422 |