INDEX

Explanations

recognition of specific media outlets and publications

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

media publication 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

헌

-0.37

 CanadaChoose

-0.35

፩

-0.34

铺

-0.33

oa̍t

-0.32

 preferable

-0.32

 equal

-0.31

日発売

-0.31

Искәрмәләр

-0.31

ThroughAttribute

-0.30

POSITIVE LOGITS

 AppDelegate

0.58

MigrationBuilder

0.54

abetes

0.52

EndInit

0.50

Jereo

0.49

pidou

0.48

Sucesor

0.48

esez

0.48

Yves

0.47



0.46

Activations Density 0.002%

recognition of specific media outlets and publications

media publication names

No Comments

No Known Activations