Neuronpedia

APISteer SAE Evals Blog/PodcastNEW Slack Privacy & Terms Contact

© Neuronpedia 2025

Privacy & Terms Blog/Podcast GitHub Slack Twitter Contact

Home
Dunefsky · Chlenski · Transcoders Enable Fine-Grained Interpretable Circuit Analysis
GPT2-Small
Transcoders Residuals
1-TRES-DC
4034

INDEX

Explanations

instances of the word "need" in various contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 rigged

-0.66

bowl

-0.66

lamm

-0.62

izon

-0.62

dylib

-0.61

creen

-0.61

��

-0.60

rams

-0.60

 Pend

-0.59

 conspiring

-0.59

POSITIVE LOGITS

lessly

0.84

ache

0.79

yle

0.71

 lear

0.70

luster

0.69

HAEL

0.67

aca

0.66

onian

0.65

lust

0.65

LESS

0.64

Activations Density 0.069%

No Known Activations