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

 luck

-0.72

 trump

-0.71

ateurs

-0.69

iary

-0.69

ements

-0.65

 crypto

-0.64

 favorites

-0.63

 outweigh

-0.62

 extras

-0.62

SIM

-0.61

POSITIVE LOGITS

idth

0.96

illard

0.82

psons

0.73

 Slug

0.72

£ı

0.68

 Seym

0.68

ili

0.67

helps

0.67

holder

0.66

igslist

0.66

Activations Density 0.000%

No Known Activations

This feature has no known activations.