INDEX

Explanations

references to ordinal positions in lists or sequences

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Configuration

ctigges/pythia-70m-deduped__res-sm_processed/1-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.1.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

MY

-1.48

 neither

-1.47

naments

-1.44

alth

-1.41

 opener

-1.41

 both

-1.35

 thirst

-1.31

 Auth

-1.31

ursed

-1.31

aven

-1.29

POSITIVE LOGITS

 ones

2.00

 dozen

1.89

enez

1.63

blic

1.61

heses

1.50

\)

1.48

half

1.46

iples

1.45

()),

1.44

hand

1.43

Activations Density 0.136%

references to ordinal positions in lists or sequences

No Comments

No Known Activations