INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 conclud

-0.77

uyomi

-0.71

zhen

-0.69

ython

-0.69

ahime

-0.68

 Angle

-0.68

igham

-0.65

zsche

-0.65

antam

-0.64

DragonMagazine

-0.63

POSITIVE LOGITS

cedented

0.74

heres

0.70

urity

0.68

hered

0.66

chairs

0.64

 recogn

0.63

IDs

0.61

"}],"

0.60

GMT

0.59

uties

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.