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

 Flickr

-0.77

byte

-0.72

 Byte

-0.64

 Chic

-0.63

FontSize

-0.63

 Likes

-0.61

Byte

-0.61

 Fenrir

-0.61

hero

-0.60

 Angry

-0.60

POSITIVE LOGITS

chio

0.83

neau

0.75

ulation

0.72

hers

0.71

enda

0.71

iliation

0.71

zza

0.66

nas

0.66

 Pavilion

0.65

 Lumpur

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.