INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

orus

-0.73

cloud

-0.69

 Bowser

-0.65

lag

-0.65

schild

-0.64

sworth

-0.64

 indexes

-0.63

ippi

-0.63

sonian

-0.62

gotten

-0.62

POSITIVE LOGITS

 enthusi

0.72

 arrang

0.69

 Flavoring

0.69

EStreamFrame

0.67

 confir

0.67

 unden

0.66

 Rehab

0.65

 proble

0.65

 comprom

0.64

 Vend

0.64

Activations Density 0.000%

No Known Activations

This feature has no known activations.