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

bia

-0.77

 sockets

-0.70

 Atom

-0.69

Rat

-0.69

 perspect

-0.68

 dwarves

-0.67

zinski

-0.66

 Afric

-0.64

wana

-0.62

 Addiction

-0.62

POSITIVE LOGITS

illac

0.72

rup

0.71

abouts

0.68

bell

0.65

quez

0.64

izon

0.64

ãģ®

0.63

 Herrera

0.63

fa

0.62

ukong

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.