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

¼

-0.71

Cos

-0.67

bro

-0.63

cut

-0.63

cos

-0.62

 loft

-0.61

Phi

-0.61

 Centauri

-0.61

Catal

-0.61

lim

-0.60

POSITIVE LOGITS

akeru

0.86

owship

0.78

yrinth

0.70

ocene

0.70

 intersections

0.69

apego

0.69

conservancy

0.68

ollar

0.67

 intersection

0.67

 sclerosis

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.