INDEX

Explanations

phrases indicating potential actions or possibilities

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

Ħ

-3.05

Ĭ

-2.89

ĳ

-2.86

»¿

-2.84

£

-2.73

±

-2.73

Ŀ

-2.69

¦

-2.69

ĸ

-2.63

¡

-2.62

POSITIVE LOGITS

 duplicate

1.73

 save

1.72

 execute

1.64

him

1.62

 lose

1.62

 devote

1.61

 avoid

1.61

rees

1.60

 realize

1.60

 eliminate

1.60

Activations Density 0.368%

phrases indicating potential actions or possibilities

No Comments

No Known Activations