INDEX

Explanations

negative statements or negations

oai_token-act-pair · gpt-3.5-turbo

the word "not" and its various uses in negating statements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

stakes

-0.73

 Circuit

-0.71

eers

-0.66

 Companies

-0.64

eur

-0.64

 Inventory

-0.63

 Tycoon

-0.62

 Contrast

-0.60

itor

-0.60

assetsadobe

-0.59

POSITIVE LOGITS

icably

1.47

icable

1.24

 necessarily

1.22

epad

1.21

orious

1.00

withstanding

0.93

ched

0.93

hin

0.90

yet

0.88

ifications

0.87

Activations Density 0.202%

negative statements or negations

the word "not" and its various uses in negating statements

No Comments

No Known Activations