INDEX

Explanations

names and surnames, particularly repeated ones

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

surname parts

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

Personendaten

-0.87

 autorytatywna

-0.79

 propOrder

-0.76

 disambiguazione

-0.73

 चीज़ों

-0.73

 contextLoads

-0.65

 للمعارف

-0.63

 Comprometido

-0.60

hyrchwyd

-0.58

----</

-0.57

POSITIVE LOGITS

cpl

0.45

UALA

0.44

 Golf

0.44

Bat

0.43

 Land

0.43

Ply

0.42

 Grid

0.42

Bet

0.42

Ent

0.41

 Landing

0.41

Activations Density 0.003%

names and surnames, particularly repeated ones

surname parts

No Comments

No Known Activations