INDEX

Explanations

quantifiable terms relating to time, duration, and conditions for transactions or events

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

 heels

-2.03

 accordingly

-1.66

 towels

-1.62

poons

-1.58

 toes

-1.51

 stride

-1.45

 blankets

-1.39

 yourselves

-1.39

 accustomed

-1.38

 closet

-1.37

POSITIVE LOGITS

 purposes

2.51

 sake

2.27

 reasons

1.90

example

1.85

 reason

1.82

got

1.67

ymp

1.64

vall

1.57

eries

1.55

better

1.54

Activations Density 2.341%

quantifiable terms relating to time, duration, and conditions for transactions or events

No Comments

No Known Activations