INDEX

Explanations

terms related to defects or defective items

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

quality defects

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

 pinulongan

-0.70

Jeografia

-0.68

 kasarigan

-0.62

-0.58

 Económica

-0.57

 Lari

-0.57

 Marius

-0.57

mari

-0.57

 Heisenberg

-0.56

mycompany

-0.56

POSITIVE LOGITS

 defect

1.12

 Defect

1.11

 defects

1.00

defect

0.95

Defect

0.95

 Defects

0.90

 defective

0.90

 Defective

0.81

 defectos

0.69

 defekt

0.67

Activations Density 0.012%

terms related to defects or defective items

quality defects

No Comments

No Known Activations