INDEX

Explanations

punctuation marks and slashes

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

scientific URLs and code

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

jenigen

-0.43

 embarazada

-0.42

 llenos

-0.41

 lleno

-0.41

 spørs

-0.40

 onderhoud

-0.39

 conocemos

-0.38

 extérieures

-0.37

 økonom

-0.37

 obě

-0.36

POSITIVE LOGITS

 Administrativna

0.65

 autorytatywna

0.64

setVerticalGroup

0.61

 المعيارى

0.60

 contextLoads

0.59

 initComponents

0.59

onViewCreated

0.57

TokenNameLBRACE

0.55

########.

0.55

ServletException

0.55

Activations Density 0.002%

punctuation marks and slashes

scientific URLs and code

No Comments

No Known Activations