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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

nect

-0.88

cius

-0.86

cil

-0.78

 Grounds

-0.76

vae

-0.75

Civ

-0.74

paio

-0.71

addons

-0.66

ypes

-0.65

renheit

-0.64

POSITIVE LOGITS

é¾

0.75

 Leviathan

0.75

 Jagu

0.68

 Archdemon

0.62

ARC

0.62

ARCH

0.62

Cost

0.61

Introduction

0.60

Shin

0.60

 scissors

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.