INDEX

Explanations

phrases indicating the start of a new segment or topic

oai_token-act-pair · gpt-3.5-turbo

the word "Here" indicating a list or detailed explanation is forthcoming

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

ONSORED

-0.62

)].

-0.60

mong

-0.60

Doctors

-0.59

 absor

-0.58

 Archdemon

-0.56

existent

-0.56

 acupuncture

-0.54

vib

-0.54

 Circle

-0.53

POSITIVE LOGITS

tics

1.29

tical

1.25

abouts

1.19

tic

1.12

ford

0.81

 Comes

0.80

after

0.77

here

0.77

yers

0.76

fore

0.76

Activations Density 0.041%

phrases indicating the start of a new segment or topic

the word "Here" indicating a list or detailed explanation is forthcoming

No Comments

No Known Activations