INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

ettings

-0.88

ritic

-0.72

thood

-0.71

chin

-0.70

ween

-0.69

aeper

-0.68

cano

-0.66

ounter

-0.64

otin

-0.64

 Pistons

-0.64

POSITIVE LOGITS

Ø¨

0.78

 Latvia

0.68

å¥³

0.66

Ø¯

0.66

Ø³

0.65

è¯

0.64

ÙĦ

0.63

 Debor

0.62

ãĥ´

0.61

INAL

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.