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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Silk

-0.73

;;;;;;;;;;;;

-0.70

lde

-0.68

arte

-0.68

osterone

-0.67

riel

-0.65

 Enhancement

-0.65

ository

-0.64

framework

-0.64

 Marketable

-0.64

POSITIVE LOGITS

opal

0.68

fam

0.65

 helicop

0.64

NY

0.61

abies

0.60

 parach

0.60

 franchises

0.59

uggage

0.59

odka

0.58

ixel

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.