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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

UE

-0.79

ODY

-0.64

 campaigner

-0.63

 rall

-0.63

 condemnation

-0.62

 corrid

-0.62

ECK

-0.61

èĥ

-0.60

loo

-0.60

 closures

-0.60

POSITIVE LOGITS

arial

0.74

zen

0.66

inqu

0.66

abis

0.65

 Scand

0.63

adjusted

0.63

teenth

0.63

comp

0.62

ibu

0.62

acial

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.