INDEX

Explanations

instances of relationships to a character's former self or past experiences

oai_token-act-pair · gpt-3.5-turbo

references to relationships or roles involving "former" and "new" statuses

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ghai

-0.87

ibaba

-0.72

gio

-0.71

orius

-0.71

]'

-0.70

 themselves

-0.69

ocene

-0.68

hovah

-0.67

inctions

-0.67

uph

-0.67

POSITIVE LOGITS

 colleague

1.37

 cousin

1.25

 partner

1.24

 girlfriend

1.23

 teammates

1.21

 fian

1.18

 wife

1.18

 teammate

1.18

 friend

1.18

 buddies

1.16

Activations Density 0.168%

instances of relationships to a character's former self or past experiences

references to relationships or roles involving "former" and "new" statuses

No Comments

No Known Activations