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

uyomi

-0.86

 occas

-0.77

çīĪ

-0.75

pse

-0.73

 showc

-0.70

spons

-0.70

 antioxid

-0.70

 Lanka

-0.69

©¶æ

-0.68

 Unsure

-0.68

POSITIVE LOGITS

erent

0.66

tom

0.66

anon

0.65

anca

0.64

drawn

0.64

anza

0.64

lement

0.63

how

0.63

remlin

0.63

 Dino

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.