INDEX

Explanations

lines of code or directives related to programming

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

ational

-2.00

oprecip

-1.58

ittal

-1.51

 puck

-1.49

undy

-1.45

apine

-1.44

onomic

-1.42

 nearby

-1.40

ney

-1.39

pere

-1.39

POSITIVE LOGITS

²

3.55

Į

3.43

½

3.29

»

3.28

´

3.22

¼

3.17

Ĥ¬

3.15

º

3.15

ĳ

3.12

ĥ½

3.11

Activations Density 0.129%

lines of code or directives related to programming

No Comments

No Known Activations