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

ounces

-0.72

cipled

-0.69

 respecting

-0.67

etheless

-0.66

onymous

-0.65

DEP

-0.62

ially

-0.61

soever

-0.61

 sorely

-0.60

FACE

-0.60

POSITIVE LOGITS

 clitor

0.77

ivas

0.70

ospels

0.68

igham

0.67

aez

0.63

 jokes

0.63

 dysph

0.62

 Porn

0.62

Untitled

0.60

 Tatt

0.60

Activations Density 0.000%

No Known Activations

This feature has no known activations.