INDEX

Explanations

references to different types of food preparation and consumption methods, specifically focusing on canned and processed foods

oai_token-act-pair · gpt-3.5-turbo

references to canned and processed food products

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

hole

-0.85

yip

-0.83

kept

-0.83

holes

-0.83

antics

-0.79

anship

-0.77

ouf

-0.77

utra

-0.76

aughty

-0.74

lust

-0.74

POSITIVE LOGITS

 goods

1.11

 versions

0.93

 vegetables

0.85

 foods

0.82

 fruits

0.80

 vegetable

0.80

 materials

0.78

ness

0.78

 meats

0.78

 beverages

0.77

Activations Density 0.157%

references to different types of food preparation and consumption methods, specifically focusing on canned and processed foods

references to canned and processed food products

No Comments

No Known Activations