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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

eral

-0.86

emetery

-0.82

erity

-0.75

amus

-0.74

agically

-0.71

ero

-0.71

ingly

-0.71

ably

-0.68

 Breach

-0.68

eem

-0.67

POSITIVE LOGITS

ĳå£«

0.74

 puff

0.68

 disapp

0.67

 Polar

0.67

 complain

0.67

ãĥ¼ãĥ³

0.64

ãĤ»

0.64

 Feedback

0.63

=]

0.63

notes

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.