INDEX

Explanations

phrases indicating categories or types of things

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

some sort of noun

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

 paciencia

-0.44

 colonnes

-0.41

 Wohnen

-0.40

 Inscrivez

-0.39

 thermique

-0.38

 cuivre

-0.38

 soledad

-0.37

 Bewertungen

-0.36

 énergétique

-0.35

 suspendu

-0.35

POSITIVE LOGITS

 somehow

0.71

 jakiś

0.69

 jakieś

0.69

GEBURTSDATUM

0.69

 Somehow

0.68

uxxxx

0.68

something

0.66

 somethin

0.65

Somehow

0.64

reportWebVitals

0.64

Activations Density 0.010%

phrases indicating categories or types of things

some sort of noun

No Comments

No Known Activations