INDEX

Explanations

references to user actions and personal experiences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 39-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_39/width_16k/average_l0_131

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.39.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

 extensively

-0.49

quickly

-0.45

 quickly

-0.45

 surla

-0.44

featureID

-0.44

 constantemente

-0.44

caref

-0.44

 quietly

-0.44

SharedCtor

-0.44

 swiftly

-0.44

POSITIVE LOGITS

AndEndTag

0.56

ใช้งาน

0.48

 antemano

0.47

ásban

0.45

 consume

0.45

duled

0.44

än

0.44

esModule

0.44

 iconLine

0.44

bibfield

0.43

Activations Density 0.703%

references to user actions and personal experiences

No Comments

No Known Activations