INDEX

Explanations

names of people and institutions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 10-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_10/width_16k/average_l0_113

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.10.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

 Monfieur

-0.79

 pleaſure

-0.72

 kaarangay

-0.68

 Wikimedijinoj

-0.65

 ſch

-0.63

IsContent

-0.62

 ſta

-0.62

 Fenster

-0.61

 Nieren

-0.60

 Verſ

-0.60

POSITIVE LOGITS

DockStyle

0.40

men

0.34

 Bonnet

0.34

initView

0.33

 Ababa

0.31

ima

0.30

<eos>

0.29

 Wives

0.29

 Bieber

0.29

 setId

0.29

Activations Density 0.745%

names of people and institutions

No Comments

No Known Activations