INDEX

Explanations

references to government officials and ministers

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

ministers and related titles

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

TagMode

-0.54

adaptiveStyles

-0.50

Демографія

-0.48

ovací

-0.47

 препратки

-0.46

archie

-0.45

رشف

-0.45

 насељу

-0.44

 Bigger

-0.42

 finer

-0.42

POSITIVE LOGITS

 Minister

0.81

 minister

0.71

Minister

0.69

 Ministry

0.65

 ministres

0.64

 ministre

0.61

 ministers

0.59

 Ministre

0.59

minister

0.58

 Ministro

0.58

Activations Density 0.020%

references to government officials and ministers

ministers and related titles

No Comments

No Known Activations