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

INESS

-0.79

Assembly

-0.70

 Became

-0.67

 Transactions

-0.65

lopp

-0.62

 Passion

-0.61

SUN

-0.60

PAT

-0.60

ocations

-0.59

hra

-0.59

POSITIVE LOGITS

zer

1.18

sub

0.89

 short

0.81

zers

0.68

icative

0.67

wra

0.63

ractor

0.62

isher

0.61

Saharan

0.60

 brief

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.