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

lihood

-0.72

epad

-0.65

ening

-0.64

 subt

-0.64

 Days

-0.63

ened

-0.63

 Months

-0.62

culation

-0.61

brance

-0.61

 Physical

-0.60

POSITIVE LOGITS

Anonymous

0.73

 eleph

0.72

âĶĢâĶĢâĶĢâĶĢ

0.68

leans

0.67

utch

0.67

illet

0.67

hello

0.66

izon

0.65

Rust

0.65

lambda

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.