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

enance

-0.74

SIGN

-0.72

una

-0.69

otte

-0.66

ãĤ¼ãĤ¦ãĤ¹

-0.63

hang

-0.63

hu

-0.60

cair

-0.59

ows

-0.59

 obstruction

-0.59

POSITIVE LOGITS

ymph

0.80

Sax

0.74

humans

0.71

Arab

0.70

maxwell

0.69

mbudsman

0.69

 =================

0.68

 Seym

0.67

++;

0.65

mbuds

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.