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

 dunno

-0.71

oun

-0.64

 shorth

-0.64

 flared

-0.62

 unres

-0.61

llor

-0.61

 darts

-0.61

 cords

-0.61

iannopoulos

-0.61

etter

-0.60

POSITIVE LOGITS

rate

0.73

Bring

0.72

Americans

0.68

Fil

0.67

Adds

0.67

Enable

0.67

edes

0.66

OG

0.64

Congratulations

0.64

idia

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.