INDEX

Explanations

positive attributes and energetic descriptions related to individuals and their actions

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

ĨĴ

-1.37

heid

-1.32

ĭ

-1.32

iscus

-1.30

³

-1.28

pshire

-1.28

isco

-1.26

icted

-1.25

ucc

-1.25

 Thames

-1.23

POSITIVE LOGITS

 emphasis

1.47

 fear

1.44

 plural

1.35

 pursuit

1.34

 part

1.28

shy

1.27

 behaviors

1.27

↵	↵

1.27

 achievement

1.25

ale

1.25

Activations Density 4.974%

positive attributes and energetic descriptions related to individuals and their actions

No Comments

No Known Activations