INDEX

Explanations

references to quotes and statements made by individuals, particularly in a professional or analytical context

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

analyst's opinion or statement

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

ExecuteReader

-0.48

 démocratique

-0.46

 gráfica

-0.41

 Actividades

-0.41

 espirituales

-0.40

Política

-0.40

 humaines

-0.38

 médicaux

-0.38

 enfermo

-0.37

 manualmente

-0.37

POSITIVE LOGITS

发表于

0.64

 informée

0.56

tagHelperRunner

0.56

MessageOf

0.53

 StatelessWidget

0.52

évaluateur

0.52

 становника

0.51

NameInMap

0.51

SpringRunner

0.45

 utom

0.44

Activations Density 0.014%

references to quotes and statements made by individuals, particularly in a professional or analytical context

analyst's opinion or statement

No Comments

No Known Activations