INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

)?

-0.66

)))

-0.63

\'

-0.62

Corp

-0.62

Ð¾

-0.62

))

-0.62

 Journals

-0.61

ell

-0.61

Ç

-0.60

Fixed

-0.60

POSITIVE LOGITS

soDeliveryDate

0.74

 airst

0.73

agra

0.70

mares

0.66

 blockade

0.65

anni

0.64

 poppy

0.63

air

0.63

ipel

0.62

azes

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.