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

sonian

-0.72

Reviewed

-0.70

 compr

-0.68

Jr

-0.65

 warr

-0.65

soever

-0.63

 Armory

-0.63

 manag

-0.63

dfx

-0.62

Magikarp

-0.62

POSITIVE LOGITS

 Braun

0.76

acers

0.69

oru

0.68

ims

0.65

insk

0.65

olen

0.65

 Kurd

0.64

 biome

0.63

Seb

0.62

 Schne

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.