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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 glim

-0.77

orem

-0.75

atana

-0.71

ascript

-0.69

dinand

-0.68

interrupted

-0.68

anmar

-0.68

ugu

-0.67

milo

-0.67

emort

-0.66

POSITIVE LOGITS

SPONSORED

0.72

 carriers

0.65

 Brewers

0.64

 Mutual

0.63

 Customs

0.62

 monop

0.62

OCD

0.61

IENT

0.60

 Sight

0.60

 VIDEOS

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.