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

ebus

-0.71

 hepat

-0.67

 poster

-0.65

 deficiency

-0.64

 protector

-0.64

 tatt

-0.63

wig

-0.63

 proficiency

-0.63

 neglig

-0.63

 remnant

-0.62

POSITIVE LOGITS

GREEN

0.78

quickShipAvailable

0.77

INFO

0.72

Joe

0.69

ilight

0.68

Smy

0.68

 Welch

0.66

elight

0.65

olla

0.65

ä

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.