Neel Nanda

Researcher whose biography, checked in October 2026, describes him as leading mechanistic interpretability at Google DeepMind.

PersonQualified confidence

Checked 2 October 2026 · AI-assisted review

Self-reported leadership of Google DeepMind’s mechanistic interpretability team as checked on 2 October 2026.

Neel Nanda — About
Self-reported leadership of Google DeepMind’s mechanistic interpretability team as checked on 2 October 2026.
Wikidata entity identity and cross-references
The entity identity and its recorded authority identifiers; not an independent authority-register validation.

What this check doesn’t establish

This is the researcher’s own biography; no employer confirmation or priority claim as a founder of the field is asserted.

Confidence applies to the scope above, not every claim made by the underlying source. How confidence works

Quoted in the essay for a judgement against his own field's most ambitious goal, which is why it carries weight there: the discount comes from someone with every incentive to talk the programme up.

In the essays

You Can't Shame a Fork
Read the essay

Sources

Neel Nanda on the limits of mechanistic interpretability
Works

Record history

Catalogue source: knowledge/people/neel-nanda.md

Generated by claude-code/claude-opus-5 on 2026-07-25