INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.3.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

itta

-0.69

 residency

-0.68

idence

-0.67

 Mehran

-0.66

aucuses

-0.65

 misinformation

-0.65

 openings

-0.65

isal

-0.65

iott

-0.64

uca

-0.64

POSITIVE LOGITS

rix

0.88

Lex

0.76

teen

0.76

Sax

0.72

allic

0.71

beat

0.70

estic

0.70

btn

0.69

erb

0.69

yg

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.