INDEX

Explanations

references to insiders or exclusive information about celebrities

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

 mécan

-0.48

 éduc

-0.47

 kapac

-0.46

actéristi

-0.46

 capacity

-0.45

 nauczy

-0.41

 proy

-0.41

 úrov

-0.40

 possibilités

-0.40

 δο

-0.40

POSITIVE LOGITS

 showbiz

0.64

 paparazzi

0.63

 celebrity

0.62

 celebs

0.59

 celeb

0.56

 Instagram

0.54

celebrity

0.52

 tabloid

0.52

 Hollywood

0.52

RegressionTest

0.49

Activations Density 0.290%

references to insiders or exclusive information about celebrities

No Comments

No Known Activations