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

stories

-0.73

 Tanz

-0.69

CHO

-0.67

 likeness

-0.64

--------------------------------------------------------

-0.63

 Synd

-0.61

 "$:/

-0.61

bott

-0.61

bots

-0.60

osen

-0.60

POSITIVE LOGITS

urgical

0.72

eport

0.72

awaru

0.70

 shone

0.70

beit

0.69

aylor

0.68

bol

0.68

kered

0.66

eval

0.66

DB

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.