INDEX

Explanations

phrases or instances related to signing up for different services, agreements, or subscriptions

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "sign."

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

ĸļ

-0.83

»Ĵ

-0.68

IRO

-0.65

âĨĳ

-0.65

ecause

-0.65

=~=~

-0.63

amily

-0.61

@#&

-0.60

 nerv

-0.59

ORGE

-0.58

POSITIVE LOGITS

atories

1.33

ificantly

1.02

atory

1.01

posts

0.93

ifying

0.86

posted

0.86

ific

0.85

alled

0.84

ificant

0.84

iff

0.82

Activations Density 0.024%

phrases or instances related to signing up for different services, agreements, or subscriptions

occurrences of the word "sign."

No Comments

No Known Activations