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

 Slay

-0.69

Flight

-0.64

 Invaders

-0.62

 Bones

-0.62

ItemImage

-0.61

 Minimum

-0.60

 rook

-0.59

catentry

-0.59

WORK

-0.58

 Cohn

-0.57

POSITIVE LOGITS

osate

1.03

xon

0.95

ofi

0.78

iott

0.76

bury

0.73

opus

0.71

itia

0.70

olver

0.68

soever

0.67

-[

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.