INDEX

Explanations

references to legal proceedings or court-related terminology

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

Ĥ

-1.79

²

-1.76

 warfare

-1.75

quo

-1.74

°

-1.68

ĸ

-1.56

pires

-1.54

ķ

-1.53

Į

-1.52

ĺ

-1.50

POSITIVE LOGITS

ycin

1.70

1.68

asp

1.60

ths

1.58

teenth

1.57

 Minutes

1.43

itars

1.42

%--

1.41

 minutes

1.41

uis

1.41

Activations Density 0.105%

references to legal proceedings or court-related terminology

No Comments

No Known Activations