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

quished

-0.78

ovie

-0.77

£ı

-0.77

ãĥ¯ãĥ³

-0.75

 Cipher

-0.74

bleacher

-0.69

 Fedora

-0.68

unin

-0.68

reau

-0.68

Downloadha

-0.67

POSITIVE LOGITS

attribute

0.75

ples

0.68

ides

0.63

angel

0.60

 affection

0.60

manship

0.58

atile

0.57

 heav

0.56

 Held

0.56

izontal

0.55

Activations Density 0.000%

No Known Activations

This feature has no known activations.