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

Madison

-0.75

Otherwise

-0.71

urous

-0.68

Downloadha

-0.67

riel

-0.66

UGE

-0.66

Assembly

-0.66

 goal

-0.64

Goal

-0.64

 Spartans

-0.64

POSITIVE LOGITS

chwitz

0.72

ande

0.72

 sshd

0.72

notations

0.71

enko

0.68

 DISTR

0.68

ussen

0.67

thora

0.63

 fluct

0.63

 misunderstand

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.