INDEX

Explanations

terms related to catastrophic or dangerous events

oai_token-act-pair · gpt-3.5-turbo

references to various objects, concepts, and events that are significant or noteworthy

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

Interstitial

-0.61

 Niet

-0.58

 thous

-0.56

 Democr

-0.55

Discussion

-0.55

 Azerb

-0.54

aeda

-0.54

\-

-0.53

enegger

-0.53

jri

-0.52

POSITIVE LOGITS

 âĢº

0.53

 truck

0.49

 eater

0.46

 robot

0.45

weed

0.44

 vill

0.44

 malt

0.43

enn

0.43

 shooter

0.43

hop

0.42

Activations Density 1.090%

terms related to catastrophic or dangerous events

references to various objects, concepts, and events that are significant or noteworthy

No Comments

No Known Activations