INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

maxwell

-0.53

NRS

-0.51

 Canaver

-0.50

theless

-0.50

Charg

-0.49

rolet

-0.48

BuyableInstoreAndOnline

-0.47

Downloadha

-0.46

interrupted

-0.46

terday

-0.46

POSITIVE LOGITS

 authent

0.45

 puppet

0.43

hij

0.41

 overth

0.40

 symptom

0.40

 sock

0.39

 homeland

0.39

 aster

0.39

 rule

0.39

 bean

0.39

Activations Density 0.000%

No Known Activations

This feature has no known activations.