© 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 · 2122 ｜ Neuronpedia

Home
GPT2-Small
Transcoders Residuals
8-TRES-DC
2122

INDEX

Explanations

the concept of acquisition or obtaining something

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

 Sahara

-0.67

��

-0.67

��

-0.64

cknow

-0.64

verson

-0.63

pei

-0.62

BLIC

-0.61

Mub

-0.61

Naz

-0.61

 Erica

-0.59

POSITIVE LOGITS

0.77

 WITHOUT

0.62

 happen

0.61

 whereas

0.60

 Airl

0.60

isin

0.59

 depend

0.59

 perm

0.59

olla

0.58

 differently

0.57

Activations Density 0.028%

No Known Activations