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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Transparency

-0.73

 tabl

-0.67

 Chop

-0.65

 disapproval

-0.65

 Vera

-0.62

 numb

-0.62

 Mansion

-0.62

 mute

-0.61

 afar

-0.60

Ivy

-0.59

POSITIVE LOGITS

ocus

0.91

iaries

0.85

retty

0.83

general

0.79

ales

0.75

ailability

0.74

 STATES

0.74

allo

0.73

iences

0.73

sein

0.72

Activations Density 0.000%

No Known Activations

This feature has no known activations.