INDEX

Explanations

commands and requests for action

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 33-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_33/width_16k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.33.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

amitié

-0.45

ต่างๆ

-0.43

ükemmel

-0.41

 various

-0.41

various

-0.40

 nemlig

-0.39

 kemungkinan

-0.39

 potentially

-0.39

 thiếu

-0.39

 çeşitli

-0.38

POSITIVE LOGITS

 fucking

0.62

RenderAtEndOf

0.61

 fuckin

0.58

 dammit

0.58

 thyself

0.58

 autorytatywna

0.57

 freakin

0.55

 goddamn

0.55

 protoimpl

0.52

Fucking

0.51

Activations Density 0.235%

commands and requests for action

No Comments

No Known Activations