INDEX

Explanations

instances of qualitative or categorical information, particularly referencing implications or claims

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

cale

-1.49

idth

-1.46

 enough

-1.43

ym

-1.37

chitz

-1.36

imus

-1.36

 Title

-1.36

acia

-1.36

ison

-1.35

documentclass

-1.35

POSITIVE LOGITS

·

3.29

¹

2.74

¿½

2.71

¶

2.65

Ĵ

2.59

Ħ

2.36

ĸ´

2.34

º

2.34

´

2.32

§

2.30

Activations Density 0.105%

instances of qualitative or categorical information, particularly referencing implications or claims

No Comments

No Known Activations