INDEX

Explanations

references to government positions and titles

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

government titles and roles

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

 Italijanski

-0.45

nloa

-0.45

 PyTuple

-0.44

 Signalez

-0.44

Vielen

-0.41

KUN

-0.41

 lingered

-0.41

 dual

-0.39

ArrowToggle

-0.39

 poisoned

-0.38

POSITIVE LOGITS

 jefe

0.59

 president

0.56

 President

0.54

president

0.54

Chief

0.54

President

0.54

 Secretary

0.54

 Bupati

0.53

Secretary

0.53

 Chief

0.53

Activations Density 0.172%

references to government positions and titles

government titles and roles

No Comments

No Known Activations