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

Negative Logits

mx

-0.80

-0.73

 Dres

-0.73

accompan

-0.72

ãĤ¦

-0.71

Wax

-0.70

ĸļ

-0.69

ãĤ¼ãĤ¦ãĤ¹

-0.68

chio

-0.68

 resil

-0.66

POSITIVE LOGITS

urry

0.72

 amplification

0.66

ivers

0.65

comed

0.65

ooter

0.64

ĺ

0.64

 Lumpur

0.64

ting

0.63

rive

0.63

 explode

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.