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

Phi

-0.76

 spice

-0.74

 spicy

-0.71

 stew

-0.69

anas

-0.67

rika

-0.67

alore

-0.66

 glac

-0.65

Ish

-0.65

 disse

-0.63

POSITIVE LOGITS

assets

0.78

film

0.73

 Unable

0.70

ciples

0.69

},{"

0.69

signed

0.69

 Accessed

0.67

letter

0.67

adjust

0.67

eous

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.