Logo Lanfrica
  • Accueil
  • Atlas
  • Analyses
  • Documentation
  • Sign in

© 2026 Lanfrica. Tous droits réservés. Tous les droits d'auteur des ressources affichées sur le site Web Lanfrica appartiennent aux détenteurs de droits d'auteur d'origine, sauf indication contraire explicite.

LSR Anchoring — Cross-Lingual Safety Steering Results

Domaine:

natural language processing

Type de record:

dataset
Créateur:
ano
Hôte:
This dataset contains all experimental results, activation caches, and run logs for the paper: Latent Space Refusal Anchoring for Low-Resource African Languages: Mechanistic Safety Recovery Without Retraining The method uses Mean Activation Steering and Sparse Autoencoder (SAE)-derived mean-activation directions (anchors) extracted from a source language (English) and applies them at

Visit

huggingface.co

Tasks

transfer learning

Languages

HausaIgboSwahiliYoruba

Tags

safetymechanistic-interpretabilityactivation-steeringmultilinguallow-resource-languagesrefusalllm-safetyafrican-languages

Licenses

mit

Similaires

LSR Anchoring Phase 2 — Cross-Lingual Safety Steering Results

LSR Anchoring Phase 2 — Cross-Lingual Safety Steering Results

This dataset contains all experimental results, activation caches, and run logs for the paper: Late