@inproceedings{yang-etal-2026-prompt,
title = "Prompt Rewrites Are Semantic Actions: Testing Whether Language Models Obey Typed Operator Laws",
author = "Yang, Zida and
Zeng, Weiqing and
You, Guquan and
Zheng, Yu",
editor = "Zhang, Yue and
Xu, Ruifeng and
Purwarianti, Ayu and
Rong, Tong and
Suzuki, Jun",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {IJCNLP}-{AACL} 2026",
month = nov,
year = "2026",
address = "Hengqin, China",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-aacl.136/",
pages = "2649--2671",
ISBN = "979-8-89176-444-6",
abstract = "Prompt rewrites can prescribe a change in an answer, not merely a change in wording. We introduce ORBIT-SIT, a framework that assigns each rewrite an explicit action on a task{'}s output space and separately measures correctness against transformed gold answers and consistency with transformed model predictions. Across eight operators, five datasets, and three instruction-tuned models, high task accuracy does not ensure compliance: Phi-4-mini achieves 90.7{\%} accuracy on SST-2 but only 16.4{\%} under a label-swap instruction. Deterministic validation and a three-annotator audit check the operator contracts. Non-deceptive binary controls retain a 17.8-percentage-point gap, and separating inference from transformation improves compliance, implicating joint execution rather than refusal alone. OrbitAlign, a quantized low-rank adaptation objective, reduces gold-grounded risk by 25.4 points on a 400-item BoolQ repair pool. The recipe lowers risk in 39 of 40 cells on two additional backbones, and same-action transfer across binary tasks reduces risk by 12.9{--}15.1 points. The binary tasks provide the primary evidence; structured tasks remain extensibility diagnostics. ORBIT-SIT makes declared prompt actions measurable and partly repairable without equating compliance with general understanding."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="yang-etal-2026-prompt">
<titleInfo>
<title>Prompt Rewrites Are Semantic Actions: Testing Whether Language Models Obey Typed Operator Laws</title>
</titleInfo>
<name type="personal">
<namePart type="given">Zida</namePart>
<namePart type="family">Yang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Weiqing</namePart>
<namePart type="family">Zeng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Guquan</namePart>
<namePart type="family">You</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yu</namePart>
<namePart type="family">Zheng</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-11</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: IJCNLP-AACL 2026</title>
</titleInfo>
<name type="personal">
<namePart type="given">Yue</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruifeng</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ayu</namePart>
<namePart type="family">Purwarianti</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Tong</namePart>
<namePart type="family">Rong</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jun</namePart>
<namePart type="family">Suzuki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Hengqin, China</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-444-6</identifier>
</relatedItem>
<abstract>Prompt rewrites can prescribe a change in an answer, not merely a change in wording. We introduce ORBIT-SIT, a framework that assigns each rewrite an explicit action on a task’s output space and separately measures correctness against transformed gold answers and consistency with transformed model predictions. Across eight operators, five datasets, and three instruction-tuned models, high task accuracy does not ensure compliance: Phi-4-mini achieves 90.7% accuracy on SST-2 but only 16.4% under a label-swap instruction. Deterministic validation and a three-annotator audit check the operator contracts. Non-deceptive binary controls retain a 17.8-percentage-point gap, and separating inference from transformation improves compliance, implicating joint execution rather than refusal alone. OrbitAlign, a quantized low-rank adaptation objective, reduces gold-grounded risk by 25.4 points on a 400-item BoolQ repair pool. The recipe lowers risk in 39 of 40 cells on two additional backbones, and same-action transfer across binary tasks reduces risk by 12.9–15.1 points. The binary tasks provide the primary evidence; structured tasks remain extensibility diagnostics. ORBIT-SIT makes declared prompt actions measurable and partly repairable without equating compliance with general understanding.</abstract>
<identifier type="citekey">yang-etal-2026-prompt</identifier>
<location>
<url>https://aclanthology.org/2026.findings-aacl.136/</url>
</location>
<part>
<date>2026-11</date>
<extent unit="page">
<start>2649</start>
<end>2671</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Prompt Rewrites Are Semantic Actions: Testing Whether Language Models Obey Typed Operator Laws
%A Yang, Zida
%A Zeng, Weiqing
%A You, Guquan
%A Zheng, Yu
%Y Zhang, Yue
%Y Xu, Ruifeng
%Y Purwarianti, Ayu
%Y Rong, Tong
%Y Suzuki, Jun
%S Findings of the Association for Computational Linguistics: IJCNLP-AACL 2026
%D 2026
%8 November
%I Association for Computational Linguistics
%C Hengqin, China
%@ 979-8-89176-444-6
%F yang-etal-2026-prompt
%X Prompt rewrites can prescribe a change in an answer, not merely a change in wording. We introduce ORBIT-SIT, a framework that assigns each rewrite an explicit action on a task’s output space and separately measures correctness against transformed gold answers and consistency with transformed model predictions. Across eight operators, five datasets, and three instruction-tuned models, high task accuracy does not ensure compliance: Phi-4-mini achieves 90.7% accuracy on SST-2 but only 16.4% under a label-swap instruction. Deterministic validation and a three-annotator audit check the operator contracts. Non-deceptive binary controls retain a 17.8-percentage-point gap, and separating inference from transformation improves compliance, implicating joint execution rather than refusal alone. OrbitAlign, a quantized low-rank adaptation objective, reduces gold-grounded risk by 25.4 points on a 400-item BoolQ repair pool. The recipe lowers risk in 39 of 40 cells on two additional backbones, and same-action transfer across binary tasks reduces risk by 12.9–15.1 points. The binary tasks provide the primary evidence; structured tasks remain extensibility diagnostics. ORBIT-SIT makes declared prompt actions measurable and partly repairable without equating compliance with general understanding.
%U https://aclanthology.org/2026.findings-aacl.136/
%P 2649-2671
Markdown (Informal)
[Prompt Rewrites Are Semantic Actions: Testing Whether Language Models Obey Typed Operator Laws](https://aclanthology.org/2026.findings-aacl.136/) (Yang et al., Findings 2026)
ACL