INDEX

Explanations

prepositions followed by nouns

oai_token-act-pair · gpt-3.5-turbo

phrases indicating quantity or amount related to actions or conditions

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

reddits

-0.84

incial

-0.75

iture

-0.66

 Cosponsors

-0.65

_-

-0.65

sit

-0.62

multi

-0.62

webkit

-0.61

 Ranking

-0.59

CLR

-0.59

POSITIVE LOGITS

 nowhere

1.28

 curiosity

1.03

 necessity

0.99

wed

0.96

 sheer

0.93

 desperation

0.93

 bounds

0.85

 boredom

0.85

 frustration

0.84

 thin

0.81

Activations Density 0.064%

prepositions followed by nouns

phrases indicating quantity or amount related to actions or conditions

No Comments

No Known Activations