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

 largeDownload

-0.78

 Harriet

-0.78

yard

-0.73

 Freed

-0.72

kees

-0.71

tumblr

-0.71

 Nurs

-0.70

 Dare

-0.70

 Unleashed

-0.70

irez

-0.70

POSITIVE LOGITS

anes

0.72

 defenses

0.69

 hubs

0.63

TBA

0.63

elin

0.61

DX

0.60

hub

0.60

comm

0.60

arr

0.59

 burgeoning

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.