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

ateur

-0.70

pie

-0.67

Skip

-0.67

ratulations

-0.67

ateurs

-0.65

pires

-0.65

 ratt

-0.62

hire

-0.61

Jump

-0.60

senal

-0.60

POSITIVE LOGITS

 Pharaoh

0.73

 Majesty

0.71

witz

0.68

 Nept

0.66

unction

0.66

 Deity

0.65

stood

0.64

 Ottoman

0.63

soever

0.63

 unaffected

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.