AI ethics evidence index Search
Source-linked daily record
Evidence record 4494 · automatically gathered

Toward Stable Value Alignment: Introducing Independent Modules for Consistent Value Guidance

Aligning large language models (LLMs) with human values typically relies on post-training or inference-time steering that directly manipulates the backbone's parameters or representation space. However, a critical gap exists: the model's residual stream is highly dynamic, in which values exist as fragile, low-dimensional properties, inherently incompatible with the stability required for consistent value expression. In this paper, we propose the Stable Value Guidance Transformer (SVGT), which ad

Record details

Published: 12 May 2026
Source: arXiv
Category: Research
Topics: Safety & alignment
Retrieved: 14 July 2026

source-onlyevidence status

These records share source-supplied organisations, an exact publisher byline, automatic topics or regions. The reason is shown on every link; related does not mean supporting, agreeing with or verifying this record.

Shared topic: Safety & alignment

SafeSteer: A Decoding-level Defense Mechanism for Multimodal Large Language Models

arXiv · 12 May 2026

Shared topic: Safety & alignment

Two Wrongs, No Right: Auditing Social-Desirability Bias in LLM Annotators for Computational Social Science

arXiv · 12 May 2026

Shared topic: Safety & alignment

Measuring What Matters Beyond Text: Evaluating Multimodal Summaries by Quality, Alignment, and Diversity

arXiv · 12 May 2026

Shared topic: Safety & alignment

Explaining and Breaking the Safety-Helpfulness Ceiling via Preference Dimensional Expansion

arXiv · 12 May 2026

Shared topic: Safety & alignment

Macro: Enhancing Multilingual Counterfactual Explanations through Alignment-as-Preference Optimization

arXiv · 12 May 2026

Shared topic: Safety & alignment

Modulation Consistency-based Contrastive Learning for Self-Supervised Automatic Modulation Classification

arXiv · 12 May 2026

How to cite this record

ethics.ai (12 May 2026), “Toward Stable Value Alignment: Introducing Independent Modules for Consistent Value Guidance,” evidence record 4494, https://ethics.ai/record/4494 (originally published by arXiv).

JSON

Use and limitations

This page is a stable index and citation surface for a source record. ethics.ai did not author the underlying report and has not independently verified every claim. Automatic topics may be imperfect. For consequential use, quote and cite the original publisher.