Modelpedia
Work in progress
About
Findings
Models
Concepts
Methods
Datasets
Sources
Light
Dark
Look Before You Leap: Universal Emergent Mechanism for Retrieval in Language Models
2025-01-22
· ICLR 2025 Poster ·
anchor
Findings
IC-372
Language models universally decompose retrieval tasks into request processing in middle layers and entity retrieval in late layers at the last token position
IC-373
In Pythia-2.8B, the specific attention heads and MLPs implementing retrieval depend on superficial input features, and request-patching preserves the natural retrieval mechanism
IC-374
Pythia models are vulnerable to prompt injection via distractor text, and request-patching from a single trusted input restores most of their accuracy