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

akeru

-0.71

interstitial

-0.71

oples

-0.70

ieri

-0.70

Downloadha

-0.69

Tip

-0.64

 graft

-0.62

£ı

-0.61

 guiActiveUn

-0.61

ometer

-0.61

POSITIVE LOGITS

noon

0.83

hello

0.75

face

0.70

FUL

0.66

sung

0.64

stellar

0.63

plot

0.63

 hello

0.62

etheus

0.62

iri

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.