INDEX

Explanations

keywords related to user interactions and ratings, probably from reviews or feedback

oai_token-act-pair · gpt-3.5-turbo

phrases related to favorites or ratings

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

aucas

-0.77

aea

-0.67

vernment

-0.66

unal

-0.64

 proced

-0.64

utical

-0.63

sheets

-0.62

lessness

-0.61

urga

-0.61

POSITIVE LOGITS

 favorite

1.36

 Favorite

1.09

favorite

0.99

Favorite

0.79

 favourite

0.78

¥µ

0.73

vertisements

0.70

BuyableInstoreAndOnline

0.70

 favorites

0.70

 Favor

0.69

Activations Density 0.025%

keywords related to user interactions and ratings, probably from reviews or feedback

phrases related to favorites or ratings

No Comments

No Known Activations