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

icult

-0.73

abet

-0.69

erent

-0.65

evil

-0.64

hazard

-0.64

igmatic

-0.62

 statewide

-0.62

yrinth

-0.62

EXP

-0.62

rawl

-0.61

POSITIVE LOGITS

 laure

0.67

 terminating

0.67

 Cance

0.66

rities

0.65

acters

0.65

 Shields

0.64

Yar

0.63

 boarding

0.63

Byr

0.62

mons

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.