© 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 · 656 ｜ Neuronpedia

Home
GPT2-Small
Transcoders Residuals
8-TRES-DC
656

INDEX

Explanations

phrases related to giving or transferring 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

ukemia

-0.79

xit

-0.75

urat

-0.73

-0.72

umption

-0.72

bians

-0.70

bf

-0.63

yll

-0.63

asis

-0.62

 fractures

-0.62

POSITIVE LOGITS

 leadership

0.76

 authority

0.75

 responsibility

0.71

 supers

0.62

 over

0.61

 gear

0.59

 agenda

0.58

endas

0.58

hon

0.58

ipment

0.58

Activations Density 0.022%

No Known Activations