INDEX

Explanations

news-related phrases and names

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

references to multimedia content such as videos, posts, and reviews

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

©¶æ¥µ

-0.63

 unforeseen

-0.58

would

-0.57

emale

-0.56

ordinate

-0.56

 perished

-0.55

ococ

-0.54

 desirable

-0.54

awar

-0.54

"]=>

-0.54

POSITIVE LOGITS

 HERE

1.29

 below

1.28

 here

1.17

 Â»

1.01

 above

1.00

below

0.96

 Below

0.91

 âĨĴ

0.87

 BELOW

0.86

 Here

0.85

Activations Density 0.243%

news-related phrases and names

references to multimedia content such as videos, posts, and reviews

No Comments

No Known Activations