INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

Fin

-0.77

 Neigh

-0.67

numbered

-0.66

 Torn

-0.65

 Lent

-0.65

Krist

-0.63

 Fernand

-0.62

ND

-0.61

Fuk

-0.61

Iss

-0.60

POSITIVE LOGITS

ulic

0.80

illac

0.79

organic

0.77

gans

0.76

uria

0.74

ocular

0.73

seless

0.73

onic

0.73

esthetic

0.72

CGI

0.70

Activations Density 0.000%

No Known Activations

This feature has no known activations.