Mechanistic interpretability
Mechanistic interpretability is losing ground in AI research: 0.7% to 0.7% of new AI papers, -0.0 points. It ranks #142 of 214 by gain in share. As of Oct 10, 2026.
- Change in share
- -0.0 pts
- Share, last 3 months
- 0.7%
- Papers, last 3 months
- 234
- Ready for products
- 1.8 / 5
More in Interpretability and mechanistic analysis
Concept based explanation0.1% → 0.1%+0.0 ptsExplanation faithfulness0.1% → 0.1%+0.0 ptsInterpretable machine learning0.1% → 0.1%+0.0 ptsHidden state probing0.0% → 0.1%0.0 ptsShortcut learning0.0% → 0.0%0.0 ptsShapley value attribution0.1% → 0.0%0.0 ptsActivation patching0.0% → 0.0%0.0 ptsRepresentation steering0.1% → 0.1%0.0 ptsFeature wise linear modulation0.0% → 0.0%0.0 ptsRepresentation engineering0.0% → 0.0%0.0 pts
Mechanistic interpretability: quick answers
No. Its share of new AI papers went from 0.7% to 0.7%, with 234 papers in the last 3 months.