INDEX

Explanations

phrases and concepts related to control and regulation, particularly in political and technological contexts

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

selection and technical terms

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.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

bootstrapcdn

-0.42

 doskon

-0.38

➌

-0.35

非常的

-0.33

 groep

-0.33

 painel

-0.33

十分的

-0.32

使其

-0.31

将其

-0.31

 operazioni

-0.31

POSITIVE LOGITS

setVerticalGroup

0.61

KURZBESCHREIBUNG

0.60

 autorytatywna

0.57

 Walkover

0.57

 مشين

0.55

Personendaten

0.55

Autoritní

0.52

esez

0.51

 noDo

0.51

彗

0.49

Activations Density 0.063%

phrases and concepts related to control and regulation, particularly in political and technological contexts

selection and technical terms

No Comments

No Known Activations