← Back to papers

Paper deep dive

Transformer Debugger

Dan Mossing, Steven Bills, Henk Tillman, Tom Dupre la Tour, Nick Cammarata, Leo Gao, Joshua Achiam, Catherine Yeh, Jan Leike, Jeff Wu, William Saunders

Year: 2024Venue: OpenAI BlogArea: Mechanistic Interp.Type: ToolEmbeddings: 0

Models: GPT-2 Small

Intelligence

Status: not_run | Model: - | Prompt: - | Confidence: 0%

Entities (0)

No extracted entities yet.

Relation Signals (0)

No relation signals yet.

Cypher Suggestions (0)

No Cypher suggestions yet.

Abstract

OpenAI's open-source interactive tool for investigating transformer internals, combining automated interpretability with sparse autoencoders for rapid model exploration.

Tags

ai-safety (imported, 100%)interpretability (suggested, 80%)mechanistic-interp (suggested, 92%)tool (suggested, 88%)

Links

Full Text

No full-text extraction is stored for this paper yet.