© 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 · 2131 ｜ Neuronpedia

Home
GPT2-Small
Transcoders Residuals
8-TRES-DC
2131

INDEX

Explanations

statements of personal opinion or expression

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

aughs

-0.78

cade

-0.71

ulence

-0.70

VW

-0.68

keyes

-0.67

":-

-0.66

agues

-0.65

\\

-0.64

eers

-0.63

lust

-0.63

POSITIVE LOGITS

 Acquisition

0.65

 receipt

0.59

 heated

0.58

 prelim

0.57

 Strateg

0.56

vironment

0.55

 Gleaming

0.55

izont

0.55

ission

0.54

uggest

0.53

Activations Density 0.065%

No Known Activations