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

 Pesh

-0.71

iston

-0.70

Pry

-0.68

apr

-0.66

imov

-0.65

 Wiley

-0.65

leton

-0.64

rieved

-0.64

ographical

-0.63

wcsstore

-0.63

POSITIVE LOGITS

ii

1.10

REAM

0.78

xual

0.68

netflix

0.67

unda

0.66

tumblr

0.66

conservancy

0.65

 incorpor

0.64

uda

0.64

 IMAGES

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.