INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

neau

-0.75

ÃįÃį

-0.71

icidal

-0.67

eteen

-0.65

itably

-0.64

ģ«

-0.64

terness

-0.64

fruit

-0.64

odynam

-0.64

livion

-0.63

POSITIVE LOGITS

msg

0.68

 certify

0.68

 reprint

0.68

ÂłÂłÂłÂł

0.66

 Creed

0.64

Ô

0.63

 Quan

0.63

 congratulate

0.62

 clip

0.62

 trademarks

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.