INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

grain

-0.76

quet

-0.70

hold

-0.64

chen

-0.63

uish

-0.63

fortune

-0.63

chens

-0.63

sole

-0.62

yo

-0.60

zed

-0.59

POSITIVE LOGITS

 incent

0.75

ICAN

0.71

ivid

0.70

DEM

0.67

WORK

0.67

BACK

0.66

Pub

0.65

assian

0.64

 elig

0.63

ADD

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.