INDEX

Explanations

verbiage related to agreements or consent with respect to receiving updates or information

oai_token-act-pair · gpt-3.5-turbo

phrases related to user options and permissions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

abal

-0.72

igans

-0.70

lled

-0.65

imer

-0.65

enhagen

-0.64

pher

-0.61

orer

-0.61

behind

-0.60

orks

-0.60

 fuss

-0.59

POSITIVE LOGITS

 Privacy

0.68

 nicotine

0.58

 donate

0.55

 conscience

0.55

{*

0.54

)=(

0.54

 Slate

0.53

license

0.53

scribe

0.52

Subscribe

0.52

Activations Density 0.026%

verbiage related to agreements or consent with respect to receiving updates or information

phrases related to user options and permissions

No Comments

No Known Activations