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

ixture

-0.71

cloud

-0.70

azaki

-0.70

ares

-0.68

arin

-0.67

archive

-0.67

ivery

-0.64

orial

-0.64

tera

-0.63

azo

-0.63

POSITIVE LOGITS

ngth

0.79

 Misty

0.76

 Morty

0.73

 Paula

0.73

ļéĨĴ

0.69

Ys

0.67

 Edited

0.67

Nad

0.65

 Clicker

0.63

Els

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.