© Neuronpedia 2026

Privacy & Terms Blog GitHub Slack Twitter Contact

Neuronpedia

Jacobian LensNEW

Natural Language

NEW Assistant AxisNEW Circuit TracerUPDATESteer SAE Evals Exports interp-engineNEWAPI Community Blog Privacy & Terms Contact

GPT2-SMALL · 8-TRES-DC · 2778 ｜ Neuronpedia

Home
GPT2-Small
Transcoders Residuals
8-TRES-DC
2778

INDEX

Explanations

elements related to identity and social roles

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Embeds

Show PlotsShow ExplanationShow ActivationsShow Test FieldShow SteerShow Link

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

SourceFile

-0.80

 largeDownload

-0.73

VERTISEMENT

-0.72

 Unloaded

-0.70

DAQ

-0.67

REDACTED

-0.65

-0.65

-0.64

FTWARE

-0.64

LOC

-0.62

POSITIVE LOGITS

 themselves

1.10

 itself

1.04

 oneself

0.88

 Yourself

0.87

 himself

0.79

 herself

0.77

antics

0.76

icates

0.76

 yourself

0.73

 Himself

0.72

Activations Density 1.855%

No Known Activations