INDEX

Explanations

qualitative descriptors and terms related to mathematical modeling and behavior

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 27-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_27/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.27.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits



-0.46

 electrolux

-0.40

 Italijanski

-0.40

 publique

-0.40

/**

-0.39

/*

-0.38

 ejecutivo

-0.37

 surla

-0.36

 IBOutlet

-0.36

POSITIVE LOGITS

tically

0.54

mente

0.52

 rzecz

0.47

的に

0.46

BagLayout

0.45

 approach

0.44

kasten

0.44

 olarak

0.44

ticamente

0.44

atically

0.44

Activations Density 1.019%

qualitative descriptors and terms related to mathematical modeling and behavior

No Known Activations