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

accompan

-0.87

theless

-0.77

obyl

-0.76

vana

-0.70

berra

-0.69

egal

-0.68

agall

-0.68

yang

-0.68

idon

-0.66

 Prelude

-0.65

POSITIVE LOGITS

ches

0.68

iles

0.64

Brand

0.64

upe

0.63

iP

0.63

 sack

0.62

istics

0.60

 Crim

0.60

Kem

0.59

itious

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.