INDEX

Explanations

expressions of gratitude and privilege

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 stratégie

-0.42

 Secretaria

-0.40

lioz

-0.40

machus

-0.39

 supaya

-0.38

 nasional

-0.38

 fatiga

-0.38

 <<<<<<<<<<<<<<

-0.36

classID

-0.36

 dagli

-0.36

POSITIVE LOGITS

 privileged

0.69

 privilege

0.64

privilege

0.64

 honored

0.64

onAttach

0.59

 الحره

0.58

 Privile

0.57

privileged

0.56

 Privilege

0.56

 privilegio

0.56

Activations Density 0.116%

expressions of gratitude and privilege

No Comments

No Known Activations