INDEX

Explanations

words related to electrical or mechanical properties and processes

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

Ķ

-1.80

ĥ

-1.73

mares

-1.60

pite

-1.59

yla

-1.58

liers

-1.58

¼

-1.56

 decision

-1.56

lier

-1.50

·¸

-1.48

POSITIVE LOGITS

operated

1.68

 PROVIDED

1.64

:#

1.63

 connected

1.56

connected

1.54

charged

1.50

electric

1.50

 wired

1.46

noreply

1.45

aeda

1.42

Activations Density 0.071%

words related to electrical or mechanical properties and processes

No Comments

No Known Activations