INDEX

Explanations

code elements related to user interface components and their properties

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

assignment operators in code

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

 يتيمه

-0.91

 صوتيه

-0.88

 للمعارف

-0.80

 المعيارى

-0.79

 betweenstory

-0.77

 autorytatywna

-0.77

<unused41>

-0.73

<unused14>

-0.73

<unused17>

-0.73

<unused28>

-0.73

POSITIVE LOGITS

was

0.45

is

0.41

IsPostBack

0.36

has

0.35

 utara

0.34

 Perubahan

0.34

 cooperativa

0.31

 cumpleaños

0.31

seguida

0.31

aksana

0.30

Activations Density 0.197%

code elements related to user interface components and their properties

assignment operators in code

No Comments

No Known Activations