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

zbek

-0.82

ohydrate

-0.73

vernment

-0.72

assisted

-0.66

 Immunity

-0.66

efficiency

-0.65

atever

-0.64

glers

-0.64

isal

-0.62

aples

-0.62

POSITIVE LOGITS

SourceFile

0.78

agonist

0.72

Interstitial

0.67

 sqor

0.67

Filename

0.64

!--

0.64

DragonMagazine

0.64

Reviewer

0.63

<[

0.62

dden

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.