INDEX

Explanations

instances where the word "not" is followed by an action or decision

oai_token-act-pair · gpt-3.5-turbo

negative statements or refusals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

krit

-0.59

eur

-0.56

 Designs

-0.55

ilo

-0.54

irlf

-0.54

Spot

-0.51

Kings

-0.51

idon

-0.50

FactoryReloaded

-0.50

 Disk

-0.50

POSITIVE LOGITS

epad

0.97

to

0.97

hin

0.95

ices

0.88

":"/

0.83

ifying

0.82

 only

0.79

ify

0.79

to

0.78

ification

0.76

Activations Density 0.039%

instances where the word "not" is followed by an action or decision

negative statements or refusals

No Comments

No Known Activations