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

llan

-0.72

SourceFile

-0.67

 euphem

-0.67

EStreamFrame

-0.66

vernment

-0.66

Arg

-0.64

nih

-0.62

 lett

-0.61

ipedia

-0.61

 plantations

-0.61

POSITIVE LOGITS

light

0.70

romeda

0.67

 Millennium

0.65

 doping

0.65

ifted

0.63

berger

0.62

agn

0.61

 TAMADRA

0.60

FSA

0.60

ãĥĵ

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.