@inproceedings{katki-etal-2025-automated,
title = "Automated Coding of Counsellor and Client Behaviours in Motivational Interviewing Transcripts: Validation and Application",
author = "Katki, Armaity and
Choi, Nathan and
Otra, Son Sophak and
Flint, George and
Zhu, Kevin and
Dev, Sunishchal",
editor = "Krishnamurthy, Parameswari and
Mujadia, Vandan and
Misra Sharma, Dipti and
Mary Thomas, Hannah",
booktitle = "NLP-AI4Health",
month = dec,
year = "2025",
address = "Mumbai, India",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2025.nlpai4health-main.4/",
pages = "25--54",
ISBN = "979-8-89176-315-9",
abstract = "Protein language models (PLMs) are powerful tools for protein engineering, but remain difficult to steer toward specific biochemical properties, where small sequence changes can affect stability or function. We adapt two prominent unsupervised editing methods: task arithmetic (TA; specifically, Forgetting via Negation) in weight space and feature editing with a sparse autoencoder (SAE) in activation space. We evaluate their effects on six biochemical properties of generations from three PLMs (ESM3, ProGen2-Large, and ProLLaMA). Across models, we observe complementary efficacies: TA more effectively controls some properties while SAE more effectively controls others. Property response patterns show some consistence across models. We suggest that the response pattern of biochemical properties should be considered when steering PLMs."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="katki-etal-2025-automated">
<titleInfo>
<title>Automated Coding of Counsellor and Client Behaviours in Motivational Interviewing Transcripts: Validation and Application</title>
</titleInfo>
<name type="personal">
<namePart type="given">Armaity</namePart>
<namePart type="family">Katki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Nathan</namePart>
<namePart type="family">Choi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Son</namePart>
<namePart type="given">Sophak</namePart>
<namePart type="family">Otra</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">George</namePart>
<namePart type="family">Flint</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Kevin</namePart>
<namePart type="family">Zhu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunishchal</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2025-12</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>NLP-AI4Health</title>
</titleInfo>
<name type="personal">
<namePart type="given">Parameswari</namePart>
<namePart type="family">Krishnamurthy</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Vandan</namePart>
<namePart type="family">Mujadia</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dipti</namePart>
<namePart type="family">Misra Sharma</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Hannah</namePart>
<namePart type="family">Mary Thomas</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Mumbai, India</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-315-9</identifier>
</relatedItem>
<abstract>Protein language models (PLMs) are powerful tools for protein engineering, but remain difficult to steer toward specific biochemical properties, where small sequence changes can affect stability or function. We adapt two prominent unsupervised editing methods: task arithmetic (TA; specifically, Forgetting via Negation) in weight space and feature editing with a sparse autoencoder (SAE) in activation space. We evaluate their effects on six biochemical properties of generations from three PLMs (ESM3, ProGen2-Large, and ProLLaMA). Across models, we observe complementary efficacies: TA more effectively controls some properties while SAE more effectively controls others. Property response patterns show some consistence across models. We suggest that the response pattern of biochemical properties should be considered when steering PLMs.</abstract>
<identifier type="citekey">katki-etal-2025-automated</identifier>
<location>
<url>https://aclanthology.org/2025.nlpai4health-main.4/</url>
</location>
<part>
<date>2025-12</date>
<extent unit="page">
<start>25</start>
<end>54</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Automated Coding of Counsellor and Client Behaviours in Motivational Interviewing Transcripts: Validation and Application
%A Katki, Armaity
%A Choi, Nathan
%A Otra, Son Sophak
%A Flint, George
%A Zhu, Kevin
%A Dev, Sunishchal
%Y Krishnamurthy, Parameswari
%Y Mujadia, Vandan
%Y Misra Sharma, Dipti
%Y Mary Thomas, Hannah
%S NLP-AI4Health
%D 2025
%8 December
%I Association for Computational Linguistics
%C Mumbai, India
%@ 979-8-89176-315-9
%F katki-etal-2025-automated
%X Protein language models (PLMs) are powerful tools for protein engineering, but remain difficult to steer toward specific biochemical properties, where small sequence changes can affect stability or function. We adapt two prominent unsupervised editing methods: task arithmetic (TA; specifically, Forgetting via Negation) in weight space and feature editing with a sparse autoencoder (SAE) in activation space. We evaluate their effects on six biochemical properties of generations from three PLMs (ESM3, ProGen2-Large, and ProLLaMA). Across models, we observe complementary efficacies: TA more effectively controls some properties while SAE more effectively controls others. Property response patterns show some consistence across models. We suggest that the response pattern of biochemical properties should be considered when steering PLMs.
%U https://aclanthology.org/2025.nlpai4health-main.4/
%P 25-54
Markdown (Informal)
[Automated Coding of Counsellor and Client Behaviours in Motivational Interviewing Transcripts: Validation and Application](https://aclanthology.org/2025.nlpai4health-main.4/) (Katki et al., NLP-AI4Health 2025)
ACL