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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

iqueness

-0.72

actionDate

-0.63

 approval

-0.60

Production

-0.59

 Manip

-0.59

 viability

-0.57

 signalling

-0.56

omorph

-0.56

 Judd

-0.55

 disapproval

-0.55

POSITIVE LOGITS

 should

1.05

should

0.94

 shouldn

0.92

 SHOULD

0.83

 ought

0.83

 book

0.79

books

0.71

book

0.71

 BOOK

0.67

Book

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.