INDEX

Explanations

terms related to norms or standards, particularly in behavioral or statistical contexts

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

 Allah

-1.80

contents

-1.47

 empowerment

-1.40

hin

-1.39

upiter

-1.38

lion

-1.37

boarding

-1.35

 chances

-1.35

vertisement

-1.34

night

-1.34

POSITIVE LOGITS

als

2.10

ativ

2.08

ality

2.06

ativity

2.03

atively

1.93

ally

1.84

iston

1.84

ative

1.84

ig

1.84

ist

1.79

Activations Density 0.228%

terms related to norms or standards, particularly in behavioral or statistical contexts

No Comments

No Known Activations