2026-09-19T00:23:16Zhttps://dspace.mit.edu/server/oai/requestoai:dspace.mit.edu:1721.1/1513452023-08-01T04:00:13Zcom_1721.1_7582com_1721.1_7581col_1721.1_131023
Andreas, Jacob
Hadfield-Menell, Dylan
Liu, Kevin
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer Science
2023-07-31T19:32:58Z
2023-07-31T19:32:58Z
2023-06
2023-06-06T16:34:59.367Z
https://hdl.handle.net/1721.1/151345
Large language models (LLMs) have been experiencing a rapid rise in utility, accessibility, and popularity, but there are still many areas in which they can improve. One such area for improvement is their truthfulness. We seek to improve the truthfulness of LLMs by probing their internal representations. We find that a linear probe on the last hidden layer representation is able to improve a model’s accuracy by reducing its confidence in incorrect answers. However, this probe is less effective at perturbing the model to change its behavior and driving the model towards correct answers.
M.Eng.
Massachusetts Institute of Technology
In Copyright - Educational Use Permitted
Copyright retained by author(s)
https://rightsstatements.org/page/InC-EDU/1.0/
Truthfulness in Large Language Models
Thesis
application/pdf
Master
Master of Engineering in Electrical Engineering and Computer Science
Publication
unknown
unknown
<Publication xmlns="https://www.openaire.eu/cerif-profile/1.1/" id="53d105eb-422a-4d6c-bd2c-81eca8cc3590">
<Type xmlns="https://www.openaire.eu/cerif-profile/vocab/COAR_Publication_Types">http://purl.org/coar/resource_type/c_1843</Type>
<Title>Truthfulness in Large Language Models</Title>
<PublishedIn>
<Publication>
</Publication>
</PublishedIn>
<PublicationDate>2023-06</PublicationDate>
<Authors>
<Author>
<DisplayName>Liu, Kevin</DisplayName>
<Affiliation>
<OrgUnit>
</OrgUnit>
</Affiliation>
</Author>
</Authors>
<Editors>
</Editors>
<Publishers>
<Publisher>
<DisplayName>Massachusetts Institute of Technology</DisplayName>
<OrgUnit />
</Publisher>
</Publishers>
<License>https://rightsstatements.org/page/InC-EDU/1.0/</License>
<Abstract>Large language models (LLMs) have been experiencing a rapid rise in utility, accessibility, and popularity, but there are still many areas in which they can improve. One such area for improvement is their truthfulness. We seek to improve the truthfulness of LLMs by probing their internal representations. We find that a linear probe on the last hidden layer representation is able to improve a model’s accuracy by reducing its confidence in incorrect answers. However, this probe is less effective at perturbing the model to change its behavior and driving the model towards correct answers.</Abstract>
<Access xmlns="http://purl.org/coar/access_right"
>
</Access>
</Publication>