INDEX

Explanations

metadata elements, specifically focusing on authorship information

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

start of user turn

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

 Lijst

-0.31

 llevo

-0.30

tidaknya

-0.29

 efectivamente

-0.29

 ponta

-0.29

 puol

-0.28

 directly

-0.28

 μὲν

-0.28

ysh

-0.28

 selectively

-0.27

POSITIVE LOGITS

 Numerade

0.90

 autorytatywna

0.83

GraphicsUnit

0.77

 AssemblyCompany

0.74

httphttps

0.73

IUrlHelper

0.69

 Administrativna

0.69

 Савезне

0.68

Derbyniad

0.68

MigrationBuilder

0.67

Activations Density 0.000%

metadata elements, specifically focusing on authorship information

start of user turn

No Comments

No Known Activations