INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ÙĴ

-0.74

 poles

-0.74

 insulated

-0.69

ema

-0.68

rug

-0.67

mob

-0.67

aday

-0.65

icket

-0.65

 bats

-0.65

 eaves

-0.64

POSITIVE LOGITS

udeau

0.78

 Understanding

0.70

href

0.68

irtual

0.67

SLI

0.66

tis

0.64

EStream

0.63

 Leilan

0.62

avez

0.62

 Blizzard

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.