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

izoph

-0.81

Reloaded

-0.78

arant

-0.77

meyer

-0.75

lon

-0.75

ope

-0.74

isi

-0.74

nec

-0.69

tnc

-0.69

iffe

-0.67

POSITIVE LOGITS

 Burger

0.70

NetMessage

0.62

 hole

0.61

 Thomson

0.61

 MARK

0.59

 merry

0.59

 errors

0.59

STE

0.58

 else

0.57

 burgers

0.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.