INDEX

Explanations

references to core concepts or elements in various 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

ieu

-1.85

 vain

-1.67

^](#

-1.64

 deceased

-1.60

 slightest

-1.60

icate

-1.58

empre

-1.55

ismiss

-1.51

ĻĤ

-1.48

ew

-1.46

POSITIVE LOGITS

doms

1.86

quisite

1.80

 concepts

1.65

ceptor

1.60

 chemistry

1.60

izable

1.59

 Features

1.57

 architecture

1.55

HEL

1.55

ceptors

1.52

Activations Density 0.145%

references to core concepts or elements in various contexts

No Comments

No Known Activations