INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

eele

-0.73

 Annie

-0.68

ipher

-0.64

 interstellar

-0.63

ylon

-0.61

 Morse

-0.60

 lyric

-0.60

lishes

-0.59

INTER

-0.59

SpaceEngineers

-0.58

POSITIVE LOGITS

)=(

0.76

Bal

0.69

ÄŁ

0.68

nings

0.67

 condol

0.67

edIn

0.67

DV

0.66

 Clubs

0.66

cham

0.63

 WhatsApp

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.