INDEX

Explanations

terms related to regulatory policies and awards in the context of performance recognition

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

New Auto-Interp

Configuration

ctigges/pythia-70m-deduped__res-sm_processed/2-res-sm

Prompts (Dashboard)

32,768 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

32,768

Data Type

torch.float32

Hook Name

blocks.2.hook_resid_post

Hook Layer

Architecture

standard

Context Size

128

Dataset

EleutherAI/the_pile_deduplicated

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 jurisdiction

-1.44

plete

-1.35

 inheritance

-1.35

@@

-1.32

atinib

-1.31

ies

-1.30

plicial

-1.29

 chance

-1.29

<!

-1.26

 estate

-1.26

POSITIVE LOGITS

ĻĤ

1.89

ľĵ

1.88

Ŀ

1.86

ĸ

1.84

ľ

1.81

¤

1.74

Ł

1.72

İ

1.71

º

1.71

ŀ

1.71

Activations Density 4.903%

terms related to regulatory policies and awards in the context of performance recognition

No Comments

No Known Activations