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

 Sets

-0.74

 congr

-0.68

 Order

-0.67

 suites

-0.65

 lucky

-0.63

die

-0.63

 outgoing

-0.62

chance

-0.60

 universes

-0.60

 Chips

-0.60

POSITIVE LOGITS

aeus

0.79

quickShipAvailable

0.79

Ïī

0.74

ãĥł

0.74

nia

0.69

EFF

0.69

 sermon

0.68

imester

0.68

docker

0.67

anza

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.