INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

oku

-0.78

 Hollow

-0.74

icum

-0.69

DS

-0.68

akis

-0.68

osure

-0.67

FS

-0.65

KY

-0.64

ion

-0.64

hiro

-0.64

POSITIVE LOGITS

EStream

0.77

assetsadobe

0.76

 manners

0.75

ateur

0.72

tymology

0.71

mingham

0.68

Minecraft

0.68

 cannabin

0.68

 antib

0.66

 arrivals

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.