INDEX

Explanations

hashtags and social media-related symbols especially those with numbers

oai_token-act-pair · gpt-3.5-turbo

hashtags used for categorization in text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

OHN

-0.68

Sov

-0.68

Â·

-0.67

soever

-0.67

âĢĳ

-0.65

otte

-0.64

merce

-0.62

 challeng

-0.61

-0.58

ccording

-0.58

POSITIVE LOGITS

3.58

2.33

(#

2.01

"#

2.00

.#

1.96

+#

1.94

##

1.88

1.78

###

1.66

 hashtag

1.65

Activations Density 0.016%

hashtags and social media-related symbols especially those with numbers

hashtags used for categorization in text

No Comments

No Known Activations

hashtags and social media-related symbols especially those with numbers

hashtags used for categorization in text

No Comments

No Known Activations