INDEX

Explanations

references to people and their roles or actions

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

people

np_acts-logits-general · gemini-2.0-flash

individuals or groups

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_16k/average_l0_88

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 asegurar

-0.41

 matéria

-0.40

 bütün

-0.40

 diğer

-0.40

지는

-0.38

 arrep

-0.36

 asegurarse

-0.35

 corregir

-0.35

 competitors

-0.35

 olvidar

-0.35

POSITIVE LOGITS

 persons

1.15

 Persons

1.13

 perfons

1.10

 person

1.09

eseorang

1.09

 people

1.07

 Person

1.04

Persons

1.02

 perſon

0.98

persons

0.98

Activations Density 0.578%

references to people and their roles or actions

people

individuals or groups

No Comments

No Known Activations