INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

Reddit

-0.88

cock

-0.79

\\\\

-0.70

Psy

-0.70

cture

-0.69

tags

-0.68

guns

-0.68

bug

-0.68

 istg

-0.67

podcast

-0.65

POSITIVE LOGITS

endish

0.69

 accompanies

0.62

 Andersen

0.61

 bargain

0.60

 havoc

0.60

 Bauer

0.59

emale

0.59

ainer

0.59

enz

0.59

 Clicker

0.58

Activations Density 0.000%

No Known Activations

This feature has no known activations.