INDEX

Explanations

specific terms related to recipes and cooking processes

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

specific entities and concepts

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

 prácti

-0.44

 ört

-0.43

 estatal

-0.42

 mandatario

-0.41

 münd

-0.41

 müh

-0.40

 traseira

-0.40

 gratuita

-0.40

readFileSync

-0.40

 państw

-0.39

POSITIVE LOGITS

 المعيارى

0.69

 surla

0.65

 esternos

0.58

 ब्रेकडाउन

0.56

 ModelExpression

0.56

 ddelweddau

0.54

 متعلقه

0.54

$_['

0.54

LookAnd

0.54

ItemBackground

0.52

Activations Density 0.112%

specific terms related to recipes and cooking processes

specific entities and concepts

No Comments

No Known Activations