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

EStreamFrame

-0.91

udeb

-0.71

phabet

-0.71

enegger

-0.70

 millenn

-0.70

ascript

-0.68

cule

-0.68

backer

-0.67

ynthesis

-0.66

 referen

-0.65

POSITIVE LOGITS

mpeg

0.75

 Physicians

0.68

 Discipline

0.68

 Medical

0.66

 Psychiatric

0.65

 Vish

0.65

Pv

0.65

Tik

0.63

 Liberties

0.63

LY

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.