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

²¾

-0.94

zar

-0.80

uler

-0.80

tz

-0.71

gaard

-0.69

lez

-0.66

quin

-0.63

ples

-0.62

tle

-0.62

ensen

-0.61

POSITIVE LOGITS

DragonMagazine

0.71

doors

0.69

mining

0.65

 Mess

0.63

afe

0.62

 spoil

0.61

 Dise

0.60

evil

0.60

 Defence

0.60

ACA

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.