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

 Ended

-0.71

odium

-0.70

 Hour

-0.67

 iPod

-0.66

 Thro

-0.65

 Halls

-0.65

 silenced

-0.64

enegger

-0.63

Geh

-0.62

peak

-0.62

POSITIVE LOGITS

externalActionCode

0.80

nat

0.78

tri

0.75

berra

0.72

ighter

0.71

bang

0.70

Rich

0.70

ignt

0.68

owicz

0.67

Fort

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.