© Neuronpedia 2026

Privacy & Terms Blog GitHub Slack Twitter Contact

Neuronpedia

Natural Language

NEW Assistant AxisNEW Circuit TracerUPDATESteer SAE Evals ExportsAPI Community Blog Privacy & Terms Contact

GPT2-SMALL · 8-TRES-DC · 202 ｜ Neuronpedia

Home
GPT2-Small
Transcoders Residuals
8-TRES-DC
202

INDEX

Explanations

expressions that emphasize the concept of "the way" something is done or happens

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

vable

-0.76

ASE

-0.72

Exception

-0.71

oother

-0.71

Single

-0.70

mi

-0.69

qualified

-0.68

rine

-0.67

ASED

-0.66

LOW

-0.65

POSITIVE LOGITS

alore

0.69

 Kali

0.66

merce

0.64

 atoms

0.62

DragonMagazine

0.61

 things

0.60

ubi

0.60

��

0.60

itism

0.60

ucer

0.59

Activations Density 0.021%

No Known Activations