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

 Rica

-0.93

rawdownloadcloneembedreportprint

-0.76

 Lanka

-0.73

Mub

-0.70

 Burma

-0.67

ILA

-0.67

ĳå£«

-0.67

 havens

-0.67

ÃĥÃĤ

-0.67

iless

-0.66

POSITIVE LOGITS

 hoop

0.84

 itch

0.70

mons

0.62

tag

0.61

ealous

0.61

tag

0.61

edient

0.59

 Totem

0.58

force

0.58

tion

0.56

Activations Density 0.000%

No Known Activations

This feature has no known activations.