Rule vs.Consequence: Dissociable Internal Representations of Moral Reasoning in LLMs
cs.CL
Submitted: 2026-10-12
Updated: 2026-10-12
Journal ref: Third Conference on Language Modeling (COLM 2026)
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Sources
- Where Norms and References Collide: Evaluating LLMs on Normative Reasoning
- The Llama 3 Herd of Models
- Measuring Pragmatic Influence in Large Language Model Instructions
- DeFrame: Debiasing Large Language Models Against Framing Effects
- Strategic Behavior of Large Language Models: Game Structure vs. Contextual Framing
- Normative Conflicts and Shallow AI Alignment
- Normative Reasoning in Large Language Models: A Comparative Benchmark from Logical and Modal Perspectives
- Steering Llama 2 via Contrastive Activation Addition
- Evaluating the Zero-shot Robustness of Instruction-tuned Language Models
- Steering Language Models With Activation Engineering
- Self-Instruct: Aligning Language Models with Self-Generated Instructions
- Qwen2.5 Technical Report
- Disentangling the Drivers of LLM Social Conformity: An Uncertainty-Moderated Dual-Process Mechanism
- Representation Engineering: A Top-Down Approach to AI Transparency
Related papers
- Exploring Solution Divergence and Its Effect on Large Language Model Problem Solving
- Ishigaki-IDS-Bench: A Benchmark for Generating Information Delivery Specification from BIM Information Requirements
- Subliminal Steering: Stronger Encoding of Hidden Signals
- MedStruct-S: A Benchmark for Key Discovery, Key-Conditioned QA and Semi-Structured Extraction from OCR Clinical Reports
- The End of Transformers? On Challenging Attention and the Rise of Sub-Quadratic Architectures
- Untangling the Mechanisms of Misleading Context in Medical Question Answering