INDEX

Explanations

names of people and their associated activities or accomplishments

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

names of people

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

 retired

-0.63

退休

-0.56

 Retired

-0.54

asanjo

-0.54

 retirement

-0.54

 ancianos

-0.54

 retiring

-0.52

EndInit

-0.52

 vectorielle

-0.50

 viejos

-0.50

POSITIVE LOGITS

 interns

0.60

Instagram

0.57

 instagram

0.54

Mateo

0.51

 vlog

0.50

 selfies

0.50

 millennials

0.49

 Millennials

0.49

Josh

0.49

 Instagram

0.49

Activations Density 0.112%

names of people and their associated activities or accomplishments

names of people

No Comments

No Known Activations