@inproceedings{kashyap-etal-2026-require,
title = "Do We Require Large Language Models To Be {HHH} Aligned For Reasoning, Safety, Value, and Pluralistic?",
author = "Kashyap, Gautam Siddharth and
Dras, Mark and
Naseem, Usman",
editor = "Lee, Hung-Yi and
Dev, Sunipa and
Li, Jessy and
Elliott, Desmond",
booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EMNLP} 2026",
month = oct,
year = "2026",
address = "Budapest, Hungary",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.findings-emnlp.2192/",
doi = "10.18653/2026.findings-emnlp.2192",
pages = "46936--46950",
ISBN = "979-8-89176-462-0",
abstract = "Alignment in Large Language Models (LLMs) studies whether model responses satisfy the Helpful, Harmless, and Honest (HHH) objectives. Imbalances across these objectives may produce divergent behavior across Reasoning, Safety, Value, and Pluralistic dimensions. Existing benchmarks including CR-LLM, SAFETUNEBENCH, VALUEBENCH, and PERSONA study these dimensions separately and capture partial aspects of HHH behavior. To study HHH behavior across overlapping dimensions, we introduce RSVP-HHH, a framework for HHH evaluation across all four dimensions. We construct RSVP-X, an English-HHH dataset containing 9,228 prompt-response pairs spanning 106 domains through Step 1 Latent Behavioral Projection, Step 2 Behavioral Region Overlap, Step 3 Density-Aware Downsampling, Step 4 Behavioral Compatibility Evaluation, and Step 5 Latent Refinement. Experimental results on RSVP-X show that TrinityX (46.6{\%}) outperforms AlignX (45.0{\%}), SAGE (49.1{\%}) outperforms VAL-BENCH (48.2{\%}) and R-Zero (47.6{\%}), DeepSeek (45.0{\%}) outperforms Qwen (43.7{\%}), and Gemini (49.6{\%}) outperforms Claude (47.7{\%}) in Avg across HHH objectives."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="kashyap-etal-2026-require">
<titleInfo>
<title>Do We Require Large Language Models To Be HHH Aligned For Reasoning, Safety, Value, and Pluralistic?</title>
</titleInfo>
<name type="personal">
<namePart type="given">Gautam</namePart>
<namePart type="given">Siddharth</namePart>
<namePart type="family">Kashyap</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Mark</namePart>
<namePart type="family">Dras</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Usman</namePart>
<namePart type="family">Naseem</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>Findings of the Association for Computational Linguistics: EMNLP 2026</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">979-8-89176-462-0</identifier>
</relatedItem>
<abstract>Alignment in Large Language Models (LLMs) studies whether model responses satisfy the Helpful, Harmless, and Honest (HHH) objectives. Imbalances across these objectives may produce divergent behavior across Reasoning, Safety, Value, and Pluralistic dimensions. Existing benchmarks including CR-LLM, SAFETUNEBENCH, VALUEBENCH, and PERSONA study these dimensions separately and capture partial aspects of HHH behavior. To study HHH behavior across overlapping dimensions, we introduce RSVP-HHH, a framework for HHH evaluation across all four dimensions. We construct RSVP-X, an English-HHH dataset containing 9,228 prompt-response pairs spanning 106 domains through Step 1 Latent Behavioral Projection, Step 2 Behavioral Region Overlap, Step 3 Density-Aware Downsampling, Step 4 Behavioral Compatibility Evaluation, and Step 5 Latent Refinement. Experimental results on RSVP-X show that TrinityX (46.6%) outperforms AlignX (45.0%), SAGE (49.1%) outperforms VAL-BENCH (48.2%) and R-Zero (47.6%), DeepSeek (45.0%) outperforms Qwen (43.7%), and Gemini (49.6%) outperforms Claude (47.7%) in Avg across HHH objectives.</abstract>
<identifier type="citekey">kashyap-etal-2026-require</identifier>
<identifier type="doi">10.18653/2026.findings-emnlp.2192</identifier>
<location>
<url>https://aclanthology.org/2026.findings-emnlp.2192/</url>
</location>
<part>
<date>2026-10</date>
<extent unit="page">
<start>46936</start>
<end>46950</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Do We Require Large Language Models To Be HHH Aligned For Reasoning, Safety, Value, and Pluralistic?
%A Kashyap, Gautam Siddharth
%A Dras, Mark
%A Naseem, Usman
%Y Lee, Hung-Yi
%Y Dev, Sunipa
%Y Li, Jessy
%Y Elliott, Desmond
%S Findings of the Association for Computational Linguistics: EMNLP 2026
%D 2026
%8 October
%I Association for Computational Linguistics
%C Budapest, Hungary
%@ 979-8-89176-462-0
%F kashyap-etal-2026-require
%X Alignment in Large Language Models (LLMs) studies whether model responses satisfy the Helpful, Harmless, and Honest (HHH) objectives. Imbalances across these objectives may produce divergent behavior across Reasoning, Safety, Value, and Pluralistic dimensions. Existing benchmarks including CR-LLM, SAFETUNEBENCH, VALUEBENCH, and PERSONA study these dimensions separately and capture partial aspects of HHH behavior. To study HHH behavior across overlapping dimensions, we introduce RSVP-HHH, a framework for HHH evaluation across all four dimensions. We construct RSVP-X, an English-HHH dataset containing 9,228 prompt-response pairs spanning 106 domains through Step 1 Latent Behavioral Projection, Step 2 Behavioral Region Overlap, Step 3 Density-Aware Downsampling, Step 4 Behavioral Compatibility Evaluation, and Step 5 Latent Refinement. Experimental results on RSVP-X show that TrinityX (46.6%) outperforms AlignX (45.0%), SAGE (49.1%) outperforms VAL-BENCH (48.2%) and R-Zero (47.6%), DeepSeek (45.0%) outperforms Qwen (43.7%), and Gemini (49.6%) outperforms Claude (47.7%) in Avg across HHH objectives.
%R 10.18653/2026.findings-emnlp.2192
%U https://aclanthology.org/2026.findings-emnlp.2192/
%U https://doi.org/10.18653/2026.findings-emnlp.2192
%P 46936-46950
Markdown (Informal)
[Do We Require Large Language Models To Be HHH Aligned For Reasoning, Safety, Value, and Pluralistic?](https://aclanthology.org/2026.findings-emnlp.2192/) (Kashyap et al., Findings 2026)
ACL