INDEX

Explanations

phrases indicating a negative statement or prohibition

oai_token-act-pair · gpt-3.5-turbo

negations and terms indicating the absence of something

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

itiz

-0.79

 Companies

-0.70

 Rhino

-0.70

AFP

-0.69

 Inventory

-0.68

ixel

-0.67

 Compass

-0.64

AFP

-0.63

former

-0.62

å¥

-0.61

POSITIVE LOGITS

 necessarily

1.20

icably

1.00

icable

0.99

epad

0.91

hin

0.84

 bothered

0.83

icing

0.80

 anymore

0.79

 gonna

0.78

ifications

0.76

Activations Density 0.225%

phrases indicating a negative statement or prohibition

negations and terms indicating the absence of something

No Comments

No Known Activations