INDEX

Explanations

specific decisions or actions taken by someone

oai_token-act-pair · gpt-3.5-turbo

instances of decision-making phrases

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

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

eries

-0.65

reen

-0.63

ILA

-0.62

VIDEO

-0.59

iddle

-0.59

outed

-0.56

esc

-0.56

ustration

-0.55

deals

-0.54

Lens

-0.54

POSITIVE LOGITS

to

1.08

 against

0.99

not

0.82

 unanimously

0.76

 upon

0.76

 that

0.75

 unilaterally

0.72

 beforehand

0.70

 nevertheless

0.69

 thereto

0.68

Activations Density 0.075%

specific decisions or actions taken by someone

instances of decision-making phrases

No Comments

No Known Activations