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

 Saud

-0.82

ented

-0.77

Ø§ÙĦ

-0.73

Ars

-0.68

thereum

-0.65

signed

-0.62

ection

-0.62

 protesting

-0.60

acebook

-0.60

vered

-0.59

POSITIVE LOGITS

igan

0.70

igans

0.69

 imperson

0.68

opl

0.68

!--

0.66

 Mechdragon

0.63

gren

0.62

deck

0.62

vill

0.62

mares

0.62

Activations Density 0.000%

No Known Activations

This feature has no known activations.