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

idated

-0.76

ç¥ŀ

-0.74

aughters

-0.74

idation

-0.73

atters

-0.72

edom

-0.69

fell

-0.68

rises

-0.68

ãĥĻ

-0.67

ailand

-0.67

POSITIVE LOGITS

 [...]

0.69

MMO

0.67

 USPS

0.67

 Etsy

0.66

 Funk

0.66

Eag

0.63

 Hick

0.62

 Monk

0.62

 ogre

0.62

Wa

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.