INDEX

Explanations

terms related to universal concepts or frameworks

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

universal

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

 Meksi

-0.53

 Cientí

-0.45

 Kompon

-0.42

 Irán

-0.38

 Fotografía

-0.38

 Sección

-0.37

 Meksika

-0.36

 Inscrivez

-0.36

IfNot

-0.36

 Turquía

-0.36

POSITIVE LOGITS

 universal

1.25

 Universal

1.18

universal

1.14

 UNIVERSAL

1.11

Universal

1.11

 universale

0.93

UNIVERS

0.87

 universality

0.86

 universally

0.84

 universel

0.82

Activations Density 0.008%

terms related to universal concepts or frameworks

universal

No Comments

No Known Activations