INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

ĸļ

-0.88

uyomi

-0.76

 unnecess

-0.74

ãĤ¦ãĤ¹

-0.68

 mattress

-0.63

 proceeds

-0.62

iatus

-0.61

phrine

-0.61

 amen

-0.60

 farewell

-0.60

POSITIVE LOGITS

utherland

0.84

holder

0.80

agascar

0.78

Pand

0.78

ãĥ¤

0.73

mercial

0.71

comings

0.70

push

0.69

ÏĢ

0.67

grab

0.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.