@inproceedings{levin-shahaf-2026-controlling,
title = "Controlling Creativity in Concept Space: A Framework for Generative Models",
author = "Levin, Dana and
Shahaf, Dafna",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.emnlp-main.1202/",
pages = "27501--27520",
ISBN = "979-8-89176-459-0",
abstract = "Generative models, such as text-to-image models, excel at taking textual prompts and producing semantically faithful outputs. However, the user{'}s ability to control the level of creativity of the output remains limited. We introduce Concept-Level Temperature (Concept-{\ensuremath{\tau}} ), a novel framework for controlling creativity in generative models.\par Unlike existing methods that operate on low-level features (tokens, noise, etc.), Concept-{\ensuremath{\tau}} operates at a higher level of abstraction; it systematically identifies decisions left underspecified by the prompt and realizes them according to a creativity parameter, creating a structured exploration space.\par In this work, we design general criteria for a creativity parameter, and implement an interpretable, model-agnostic method to satisfy them. We demonstrate our ideas on the text-to-image domain; empirical evaluation, both qualitative and quantitative, shows that Concept- significantly outperforms baseline methods in creative modulation and enhancement."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="levin-shahaf-2026-controlling">
<titleInfo>
<title>Controlling Creativity in Concept Space: A Framework for Generative Models</title>
</titleInfo>
<name type="personal">
<namePart type="given">Dana</namePart>
<namePart type="family">Levin</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Dafna</namePart>
<namePart type="family">Shahaf</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-10</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing</title>
</titleInfo>
<name type="personal">
<namePart type="given">Hung-Yi</namePart>
<namePart type="family">Lee</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sunipa</namePart>
<namePart type="family">Dev</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jessy</namePart>
<namePart type="family">Li</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Desmond</namePart>
<namePart type="family">Elliott</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Budapest, Hungary</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-459-0</identifier>
</relatedItem>
<abstract>Generative models, such as text-to-image models, excel at taking textual prompts and producing semantically faithful outputs. However, the user’s ability to control the level of creativity of the output remains limited. We introduce Concept-Level Temperature (Concept-\ensuremathτ ), a novel framework for controlling creativity in generative models.\par Unlike existing methods that operate on low-level features (tokens, noise, etc.), Concept-\ensuremathτ operates at a higher level of abstraction; it systematically identifies decisions left underspecified by the prompt and realizes them according to a creativity parameter, creating a structured exploration space.\par In this work, we design general criteria for a creativity parameter, and implement an interpretable, model-agnostic method to satisfy them. We demonstrate our ideas on the text-to-image domain; empirical evaluation, both qualitative and quantitative, shows that Concept- significantly outperforms baseline methods in creative modulation and enhancement.</abstract>
<identifier type="citekey">levin-shahaf-2026-controlling</identifier>
<location>
<url>https://aclanthology.org/2026.emnlp-main.1202/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>27501</start>
<end>27520</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Controlling Creativity in Concept Space: A Framework for Generative Models
%A Levin, Dana
%A Shahaf, Dafna
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-459-0
%F levin-shahaf-2026-controlling
%X Generative models, such as text-to-image models, excel at taking textual prompts and producing semantically faithful outputs. However, the user’s ability to control the level of creativity of the output remains limited. We introduce Concept-Level Temperature (Concept-\ensuremathτ ), a novel framework for controlling creativity in generative models.\par Unlike existing methods that operate on low-level features (tokens, noise, etc.), Concept-\ensuremathτ operates at a higher level of abstraction; it systematically identifies decisions left underspecified by the prompt and realizes them according to a creativity parameter, creating a structured exploration space.\par In this work, we design general criteria for a creativity parameter, and implement an interpretable, model-agnostic method to satisfy them. We demonstrate our ideas on the text-to-image domain; empirical evaluation, both qualitative and quantitative, shows that Concept- significantly outperforms baseline methods in creative modulation and enhancement.
%U https://aclanthology.org/2026.emnlp-main.1202/
%P 27501-27520
Markdown (Informal)
[Controlling Creativity in Concept Space: A Framework for Generative Models](https://aclanthology.org/2026.emnlp-main.1202/) (Levin & Shahaf, EMNLP 2026)
ACL