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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

arna

-0.71

gain

-0.68

 Karma

-0.68

 Ascension

-0.67

ufact

-0.67

Cav

-0.66

 Pact

-0.65

 Entry

-0.64

dule

-0.63

Suc

-0.63

POSITIVE LOGITS

EngineDebug

0.75

 WATCHED

0.71

Ã¨re

0.70

ently

0.70

 WARN

0.70

oult

0.69

feld

0.68

@#&

0.68

lie

0.67

udi

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.