INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

rouse

-0.91

ierrez

-0.82

kai

-0.75

Office

-0.71

Tokens

-0.70

 Mandarin

-0.69

ulhu

-0.69

ACS

-0.69

awei

-0.68

 Bonds

-0.67

POSITIVE LOGITS

 reluct

0.76

 rupt

0.70

ergus

0.67

hood

0.65

 inconsist

0.64

 disapp

0.64

cloth

0.63

issance

0.63

 deleting

0.63

 aback

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.