INDEX

Explanations

references to specific models, tests, and groups in a structured format

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

figures, models, equations, and nodes

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

GTCX

-0.81

 protoimpl

-0.80

 kaarangay

-0.79

 Numerade

-0.77

parsedMessage

-0.73

 kasarigan

-0.71

Diwedd

-0.68

 للمعارف

-0.67

 المعيارى

-0.66

+#+

-0.66

POSITIVE LOGITS

 Baus

0.41

들의

0.36

gangen

0.35

 vescovo

0.34

 lainnya

0.34

esterno

0.34

0.33

 çalışan

0.33

DJ

0.33

 Cardinal

0.32

Activations Density 0.079%

references to specific models, tests, and groups in a structured format

figures, models, equations, and nodes

No Comments

No Known Activations

references to specific models, tests, and groups in a structured format

figures, models, equations, and nodes

No Comments

No Known Activations