INDEX

Explanations

textual elements related to different writing systems and representation of languages

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

alphabets and characters

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

AddTagHelper

-0.52

StoryboardSegue

-0.50

ⓧ

-0.48

fitrión

-0.46

 Rossini

-0.43

zzleHttp

-0.43

viewWillAppear

-0.42

NavController

-0.42

 Bowen

-0.41

 optimizing

-0.41

POSITIVE LOGITS

Autoritní

0.59

 Administrativna

0.54

Referencies

0.51

Билгалдахарш

0.48

MigrationBuilder

0.47

WithIOException

0.44

Cyfeiriadau

0.42

ArrowToggle

0.41

 Мексичка

0.41

 tryck

0.41

Activations Density 0.012%

textual elements related to different writing systems and representation of languages

alphabets and characters

No Comments

No Known Activations