INDEX

Explanations

actions or commands that involve some level of authority or decision-making power

oai_token-act-pair · gpt-3.5-turbo

actions or processes that involve engagement or connection

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

quartered

-0.55

 wont

-0.50

cour

-0.49

\":

-0.48

DragonMagazine

-0.48

pmwiki

-0.47

Downloadha

-0.46

@#&

-0.45

displayText

-0.45

iae

-0.44

POSITIVE LOGITS

 oneself

0.91

uate

0.83

enance

0.82

 them

0.79

ulate

0.74

 yourself

0.70

him

0.69

urate

0.68

 ourselves

0.64

igate

0.63

Activations Density 0.631%

actions or commands that involve some level of authority or decision-making power

actions or processes that involve engagement or connection

No Comments

No Known Activations