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

heed

-0.76

itol

-0.71

 surname

-0.71

heric

-0.70

awi

-0.68

inx

-0.67

riot

-0.66

igon

-0.65

socket

-0.65

 Mirror

-0.64

POSITIVE LOGITS

ONSORED

0.73

LAB

0.66

 Advis

0.65

ecd

0.65

ados

0.63

Business

0.63

EB

0.63

 surveys

0.62

natureconservancy

0.62

ECO

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.