INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

vre

-0.68

venge

-0.68

 Mori

-0.64

 Toledo

-0.63

 Klingon

-0.63

ige

-0.63

gas

-0.62

 Hunts

-0.62

 Crimea

-0.60

 Persian

-0.60

POSITIVE LOGITS

ĸļ

0.80

idon

0.78

00200000

0.75

thora

0.72

jri

0.71

anish

0.70

poons

0.69

alsa

0.66

ijn

0.66

acly

0.65

Activations Density 0.000%

No Known Activations

This feature has no known activations.