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

utic

-0.65

ancial

-0.64

 Armageddon

-0.62

 Exploration

-0.61

cone

-0.61

ados

-0.59

 Crusade

-0.58

amiya

-0.57

 Fant

-0.57

 Apocalypse

-0.57

POSITIVE LOGITS

EStream

0.80

glers

0.69

milo

0.68

ibi

0.66

lich

0.66

----

0.65

fusc

0.64

abul

0.63

weet

0.63

unes

0.61

Activations Density 0.000%

No Known Activations

This feature has no known activations.