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

alist

-0.75

 Mandela

-0.71

 hinges

-0.70

 Gutenberg

-0.68

inness

-0.67

minist

-0.67

ulic

-0.63

 staples

-0.63

 Haram

-0.62

 upside

-0.62

POSITIVE LOGITS

thood

0.74

rollers

0.74

friend

0.71

Dispatch

0.68

eve

0.67

othes

0.66

CRIPTION

0.64

ÙĴ

0.63

perse

0.63

licts

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.