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

chery

-0.66

wide

-0.66

corn

-0.66

Enlarge

-0.64

yards

-0.63

Pak

-0.63

yard

-0.62

flush

-0.61

yah

-0.61

age

-0.61

POSITIVE LOGITS

ivable

0.76

DonaldTrump

0.71

©¶æ

0.71

etheless

0.69

 inclination

0.69

fff

0.69

eatures

0.69

odic

0.69

 appre

0.69

 Parables

0.68

Activations Density 0.000%

No Known Activations

This feature has no known activations.