INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 deposition

-0.75

 cloning

-0.74

 whisper

-0.71

 composition

-0.69

 clout

-0.66

 cascade

-0.64

umbnails

-0.63

NetMessage

-0.63

 mosqu

-0.61

iping

-0.61

POSITIVE LOGITS

acebook

0.91

ilton

0.78

retty

0.73

 sund

0.72

odd

0.69

esa

0.69

uka

0.68

starter

0.66

endment

0.65

ulations

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.