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

 Catalan

-0.67

Hug

-0.62

 French

-0.59

 happy

-0.59

aph

-0.59

 word

-0.58

Synopsis

-0.57

 french

-0.57

 lucid

-0.57

Eur

-0.57

POSITIVE LOGITS

anwhile

0.82

otype

0.79

ums

0.74

etheless

0.71

yrinth

0.70

 sett

0.68

ocaust

0.68

CLASSIFIED

0.68

udes

0.67

aspers

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.