INDEX

Explanations

words related to entitlement or privileges

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

poons

-1.94

backs

-1.67

unnable

-1.61

ontal

-1.47

 throat

-1.45

quin

-1.42

 else

-1.42

fully

-1.40

put

-1.39

xit

-1.39

POSITIVE LOGITS

ourage

1.87

Ŀ

1.76

ele

1.53

rants

1.51

ibase

1.49

angle

1.46

transcript

1.42

iendo

1.39

 bloc

1.38

array

1.38

Activations Density 0.080%

words related to entitlement or privileges

No Comments

No Known Activations