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

bold

-0.71

paren

-0.71

urst

-0.65

Pa

-0.65

started

-0.64

shut

-0.64

 Rivals

-0.64

FIR

-0.63

HCR

-0.63

ghazi

-0.62

POSITIVE LOGITS

soever

0.71

ieties

0.71

 Jinn

0.70

ollah

0.69

 Enix

0.68

 handy

0.67

 Mehran

0.66

anga

0.66

 thous

0.66

ĸļ

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.