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

fred

-0.81

ework

-0.76

eworks

-0.74

abouts

-0.71

ings

-0.69

href

-0.68

itance

-0.67

onics

-0.67

feld

-0.67

aic

-0.65

POSITIVE LOGITS

razil

0.80

muc

0.70

 Hebdo

0.67

undy

0.65

udeau

0.59

addon

0.58

 demographics

0.58

 normalized

0.57

 addon

0.57

 mirror

0.56

Activations Density 0.000%

No Known Activations

This feature has no known activations.