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

bons

-0.70

bin

-0.64

 chairs

-0.60

 illusions

-0.60

 Eisenhower

-0.59

wo

-0.59

 Bourbon

-0.58

 boredom

-0.58

Blade

-0.57

 owners

-0.57

POSITIVE LOGITS

âĢ

1.30

 âĢ

0.86

ï¸ı

0.83

 ðŁĳ

0.82

pring

0.80

 âĺ

0.78

conservancy

0.75

âĨĳ

0.74

âĢł

0.74

âľ

0.73

Activations Density 0.000%

No Known Activations

This feature has no known activations.