INDEX

Explanations

tech-related or programming terms and concepts

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

:`

-1.88

iances

-1.63

 formats

-1.54

endas

-1.50

 basis

-1.43

 purposes

-1.42

 minded

-1.39

wegian

-1.38

icum

-1.37

ictional

-1.37

POSITIVE LOGITS

ĻĤ

2.06

¿½

1.94

¦

1.91

¯

1.85

«

1.83

»

1.78

ĳ

1.72

Ĥ

1.71

Ĺ

1.62

Ĥ¬

1.61

Activations Density 5.451%

tech-related or programming terms and concepts

No Comments

No Known Activations