INDEX

Explanations

phrases related to editorial content selections

oai_token-act-pair · gpt-3.5-turbo

references to "Picks," "Articles," or similarly categorized selections

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

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

lia

-0.81

past

-0.74

stood

-0.70

gal

-0.68

bred

-0.66

igi

-0.65

romy

-0.65

bent

-0.65

acion

-0.65

gur

-0.63

POSITIVE LOGITS

 Picks

1.38

ettings

0.76

 Continue

0.75

ļéĨĴ

0.72

 Morning

0.71

 Lists

0.71

ullivan

0.70

 Annotations

0.70

 Cosponsors

0.68

 Markets

0.68

Activations Density 0.014%

phrases related to editorial content selections

references to "Picks," "Articles," or similarly categorized selections

No Comments

No Known Activations