INDEX

Explanations

terms indicating a comparison or relationship between two or more entities

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

both or either

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

έ

-0.39

 épaules

-0.39

 nadie

-0.38

POLYGON

-0.38

 genoux

-0.37

 Inscrivez

-0.36

 vastaan

-0.35

 cœurs

-0.34

vœ

-0.34

 honneur

-0.34

POSITIVE LOGITS

 Мексичка

0.90

IVEREF

0.86

تقاوى

0.79

RegistryLite

0.77

 Chwiliwch

0.75

__(/*!

0.74

owohl

0.73

 ویکی‌پدی

0.72

 both

0.71

 BOTH

0.71

Activations Density 0.038%

terms indicating a comparison or relationship between two or more entities

both or either

No Comments

No Known Activations