INDEX

Explanations

references to news articles and information

oai_token-act-pair · gpt-3.5-turbo

sequences of slashes or dividers often associated with lists or sections within text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

terday

-0.71

idious

-0.69

abies

-0.66

riages

-0.65

eryl

-0.65

pter

-0.65

nesses

-0.63

ockets

-0.62

 subsequ

-0.62

repre

-0.62

POSITIVE LOGITS

ËĪ

1.41

Via

0.98

usr

0.85

 Shutterstock

0.84

Ħ¢

0.72

————————

0.71

 BuzzFeed

0.70

 Tags

0.69

RTX

0.69

¥µ

0.67

Activations Density 0.018%

references to news articles and information

sequences of slashes or dividers often associated with lists or sections within text

No Comments

No Known Activations