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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Pavilion

-0.74

 simultane

-0.73

Flo

-0.71

 Falls

-0.70

agate

-0.65

 Ukrain

-0.65

 Gorge

-0.63

 defic

-0.63

 Unic

-0.63

 Dupl

-0.62

POSITIVE LOGITS

ortium

0.88

emonic

0.76

APD

0.75

osterone

0.71

gaard

0.71

ËĪ

0.68

USS

0.67

insula

0.67

CTV

0.67

LCS

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.