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

prints

-0.94

ãĤ©

-0.90

hower

-0.87

lease

-0.84

*/(

-0.84

ocene

-0.83

 ILCS

-0.78

cone

-0.77

isol

-0.77

elaide

-0.75

POSITIVE LOGITS

 Yemeni

0.70

MP

0.69

Maz

0.69

TC

0.67

Luc

0.65

 icing

0.64

kn

0.63

ted

0.62

 poppy

0.61

ast

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.