INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

chu

-0.88

âĺħ

-0.81

vel

-0.73

ãĢĲ

-0.72

Â¶

-0.70

å£«

-0.69

ince

-0.68

*,

-0.66

âĢł

-0.64

âĪĴ

-0.63

POSITIVE LOGITS

utterstock

0.82

ickr

0.78

 pigeon

0.73

phia

0.71

ascript

0.68

 pictured

0.66

 misunder

0.66

 overloaded

0.65

 suspended

0.64

 vetoed

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.