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

eties

-0.79

termination

-0.77

ibaba

-0.72

rican

-0.70

ocial

-0.70

ingly

-0.69

ega

-0.69

ensible

-0.68

endish

-0.67

acan

-0.67

POSITIVE LOGITS

 REPL

0.70

DIR

0.66

 Unlimited

0.65

 Explorer

0.63

 convert

0.62

 converted

0.60

});

0.59

ãĤ©

0.59

 receives

0.59

--+

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.