© Neuronpedia 2026

Privacy & Terms Blog GitHub Slack Twitter Contact

Neuronpedia

APIAssistant AxisNEW Circuit TracerNEW Steer SAE Evals Exports Community Blog Privacy & Terms Contact

GPT2-SMALL · 8-TRES-DC · 907 ｜ Neuronpedia

Home
GPT2-Small
Transcoders Residuals
8-TRES-DC
907

INDEX

Explanations

variations of the word "occupy" and concepts related to occupation or activity

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

RAW

-0.65

 antioxid

-0.62

 Shed

-0.62

 rested

-0.61

 challeng

-0.61

 Exhibit

-0.61

\\\\\\\\

-0.61

 notor

-0.60

byss

-0.60

 housed

-0.59

POSITIVE LOGITS

ancies

1.22

ancy

1.16

ances

1.16

ateurs

1.11

ations

1.07

ents

1.04

ants

1.04

ence

1.04

ators

1.04

ements

1.03

Activations Density 0.164%

No Known Activations