INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ersed

-0.74

 Admission

-0.64

shared

-0.63

drops

-0.63

ield

-0.63

DL

-0.63

pers

-0.62

lic

-0.62

hiro

-0.61

mans

-0.59

POSITIVE LOGITS

ĪĴ

0.94

uyomi

0.84

ģĸ

0.80

ŃĶ

0.78

abwe

0.77

myra

0.76

ħĭ

0.73

âĢ¢âĢ¢

0.73

jri

0.72

Ń·

0.70

Activations Density 0.000%

No Known Activations

This feature has no known activations.