INDEX

Explanations

phrases related to making and implementing changes

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

¸

-3.48

·

-3.31

Īĺ

-3.18

Ļ

-3.17

±

-3.16

ĳ

-3.06

ī

-3.04

ĥ½

-3.02

İ

-3.02

ħ

-2.98

POSITIVE LOGITS

 himself

1.96

 yourself

1.84

 ourselves

1.75

 themselves

1.74

(:

1.68

 herself

1.67

leine

1.63

 yourselves

1.61

amin

1.59

 accordingly

1.54

Activations Density 0.603%

phrases related to making and implementing changes

No Comments

No Known Activations