Entity · dataset

SH-Detection

datasetactiveprovisionalsh-detection-9967a81f·1 events·first seen 3d ago

Aliases: SH-Detection

Co-occurring entities

X-Sensitive Gemma-3-4B-IT Analysing Self-Harm Representations in Language Models: a Cross-Architecture Study

More like this (12)

SV-Detect CH-SIMS stance detection HPDv2 WaveDetect HPDv3 Safety Detection Classifier LC-DP HITL-D SHAP Grad Detect hate speech detection

Recent events (1)

5arXiv · cs.CL·3d ago·source ↗

Cross-architecture study of how LLMs internally represent self-harm content

Researchers analyze how four LLMs encode self-harm content using linear probes and contrastive direction extraction across two datasets (X-Sensitive and SH-Detection). A key finding is that self-harm information crystallizes in the final 3–7% of network layers (93–97% depth), and that probe accuracy does not correlate straightforwardly with linear separability. Gemma-3-4B shows a notably different internal representation of the contrastive self-harm direction compared to other models, with implications for detection, intervention, and LLM governance.

Evaluation and Benchmarking AI Safety Research X-Sensitive Gemma-3-4B-IT SH-Detection +1 more