INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

fters

-0.77

ansion

-0.75

eli

-0.75

fts

-0.73

atures

-0.73

athed

-0.72

agame

-0.72

acs

-0.72

uth

-0.71

SET

-0.71

POSITIVE LOGITS

 Kuro

0.71

 Ariel

0.69

Hou

0.69

 Typhoon

0.66

tru

0.65

Bye

0.64

WP

0.63

 grou

0.61

Eh

0.61

 Freddie

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.