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

icably

-0.85

xon

-0.75

ioxid

-0.71

mint

-0.70

ophobia

-0.69

etheless

-0.69

ividually

-0.68

flame

-0.68

 Frieza

-0.68

cheon

-0.67

POSITIVE LOGITS

 Kick

0.74

 Tracks

0.72

Mat

0.69

 Email

0.67

 Checks

0.67

Sak

0.65

send

0.61

 Kali

0.61

STL

0.60

 Belgium

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.