INDEX

Explanations

references to specific programming or technical concepts and actions

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

initialize your, function, achieved, Civi

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

-0.57

 continúas

-0.45

 Autorisations

-0.44

 تضيفلها

-0.43

 secador

-0.43

 кӀош

-0.41

 desmotivaciones

-0.40

 collè

-0.40

 costes

-0.40

 jurk

-0.40

POSITIVE LOGITS

 autorytatywna

0.54

 препратки

0.48

EDEFAULT

0.41

serviceWorker

0.40

 BrowserModule

0.40

 Mods

0.40

BeginContext

0.40

信

0.40

oredCriteria

0.39

 صوتيه

0.39

Activations Density 0.298%

references to specific programming or technical concepts and actions

initialize your, function, achieved, Civi

No Comments

No Known Activations