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

Reloaded

-0.77

throp

-0.67

erc

-0.66

 XIII

-0.65

iston

-0.63

stocks

-0.63

XIV

-0.62

BACK

-0.62

 exting

-0.60

 Yugoslav

-0.60

POSITIVE LOGITS

Ĥİ

0.68

gregation

0.67

fusc

0.64

NetMessage

0.64

¶æ

0.63

olated

0.61

 tagging

0.60

Lt

0.60

ffic

0.59

Saf

0.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.