INDEX

Explanations

references to the concept of "privacy" or "confidentiality"

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Configuration

ctigges/pythia-70m-deduped__mlp-sm_processed/2-mlp-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_mlp_out

Hook Layer

Architecture

standard

Context Size

128

Dataset

EleutherAI/the_pile_deduplicated

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

mann

-1.67

garten

-1.61

van

-1.59

enance

-1.53

thor

-1.50

liness

-1.43

 rhetoric

-1.42

Sr

-1.39

eman

-1.38

INGS

-1.38

POSITIVE LOGITS

"){

1.85

lickr

1.66

fiddle

1.51

optic

1.50

etica

1.48

{$

1.47

irez

1.42

Īĺ

1.40

ĻĤ

1.39

quartile

1.37

Activations Density 0.252%

references to the concept of "privacy" or "confidentiality"

No Known Activations