INDEX

Explanations

references to organizational roles and titles

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

holding positions or serving

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

 հղումներ

-0.58

GEBURTSDATUM

-0.57

Hauptartikel

-0.53

 disambiguazione

-0.51

 Италијани

-0.50

KURZBESCHREIBUNG

-0.49

دانشنامهٔ

-0.48

rungsseite

-0.46

 ब्रेकडाउन

-0.46

 nonUne

-0.45

POSITIVE LOGITS

 committee

0.47

ställ

0.43

ándola

0.43

 leadership

0.42

 exporta

0.42

 iesp

0.42

 périodes

0.42

Twice

0.40

 longtemps

0.40

heits

0.40

Activations Density 0.012%

references to organizational roles and titles

holding positions or serving

No Comments

No Known Activations