INDEX

Explanations

phrases expressing strong intentions or obligations

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

New Auto-Interp

Configuration

ctigges/pythia-70m-deduped__mlp-sm_processed/2-mlp-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_mlp_out

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

¿½

-2.14

ĥ

-1.96

±

-1.85

ħ

-1.85

³

-1.83

İ

-1.82

º

-1.81

¨

-1.77

Ħ

-1.73

Ŀ

-1.73

POSITIVE LOGITS

 downloaded

1.62

carrier

1.60

matched

1.58

earch

1.51

itzer

1.46

Ãªme

1.45

hower

1.40

 licensed

1.39

 included

1.38

 purchased

1.37

Activations Density 0.256%

phrases expressing strong intentions or obligations

No Comments

No Known Activations