INDEX

Explanations

elements related to transformation and disguise, particularly in the context of identity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 11-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_11/width_16k/average_l0_118

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.11.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

tvguidetime

-0.51

 utafitiHapana

-0.42

 Rücks

-0.41

 gifts

-0.41

 gift

-0.41

cargas

-0.39

 insuffisamment

-0.39

 Gift

-0.39

gif

-0.38

Gift

-0.38

POSITIVE LOGITS

 imperson

0.64

 disguise

0.58

 pretending

0.53

guisement

0.51

 IDENTITY

0.51

 guise

0.49

 Immer

0.47

ContentAsync

0.47

Immer

0.47

AndroidJUnit

0.47

Activations Density 0.688%

elements related to transformation and disguise, particularly in the context of identity

No Comments

No Known Activations