INDEX

Explanations

instances of categorized information or classifications

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

Category names

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

cuál

-0.38

Ingrédients

-0.35

Israël

-0.32

 poitrine

-0.31

 spot

-0.30

 character

-0.29

enx

-0.29

的地

-0.29

Quiénes

-0.28

msleep

-0.28

POSITIVE LOGITS

 للاسماء

1.17

 صوتيه

0.79

Autoritní

0.78

 محفوظة

0.71

SerializedSize

0.66

 BoxDecoration

0.66

ScopeManager

0.66

 pinulongan

0.65

Datuak

0.64

 للمعارف

0.64

Activations Density 0.032%

instances of categorized information or classifications

Category names

No Comments

No Known Activations