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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

innon

-0.78

bots

-0.77

hov

-0.74

makers

-0.72

Ö¼

-0.70

Ð´

-0.67

arnaev

-0.66

foreseen

-0.66

few

-0.66

ontent

-0.65

POSITIVE LOGITS

Yor

0.75

 Issue

0.68

 Keyboard

0.67

 relic

0.67

 grant

0.66

 writ

0.66

OP

0.65

 purse

0.65

 Newsletter

0.64

 Donation

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.