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

âĹı

-0.70

uberty

-0.66

 Mald

-0.63

okÃ©

-0.63

ossibility

-0.62

iii

-0.62

arent

-0.62

 proble

-0.62

Synopsis

-0.61

eatures

-0.60

POSITIVE LOGITS

 certs

0.80

IRT

0.71

Rx

0.70

 istg

0.68

-->

0.66

plane

0.65

raints

0.64

OTA

0.63

 Seah

0.63

plate

0.63

Activations Density 0.000%

No Known Activations

This feature has no known activations.