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

ndra

-0.75

 Debor

-0.75

enfranch

-0.73

cot

-0.73

soever

-0.72

chool

-0.71

minist

-0.71

agents

-0.67

aez

-0.66

itect

-0.66

POSITIVE LOGITS

INS

0.78

ued

0.75

 brace

0.70

osite

0.64

xes

0.63

 ranges

0.63

 conceal

0.61

 insertion

0.61

aceous

0.60

 pseudonym

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.