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

ifiable

-0.68

ç¥ŀ

-0.64

onga

-0.63

 leverage

-0.63

ko

-0.63

 repositories

-0.61

drop

-0.61

South

-0.61

Updated

-0.60

POSITIVE LOGITS

byss

0.82

unny

0.79

dylib

0.72

 Apostles

0.71

ngth

0.70

 Merry

0.64

accompan

0.64

angel

0.64

 mids

0.63

hetics

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.