INDEX

Explanations

references to notable figures and scholars in various disciplines

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

names of people

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

 wireType

-0.46

 escenas

-0.42

RegressionTest

-0.42

 facilidad

-0.40

➌

-0.40

 gancho

-0.40

 redacción

-0.39

 gravedad

-0.39

 orgullo

-0.39

 Komunikasi

-0.39

POSITIVE LOGITS

 kasarigan

0.54

脚注の使い方

0.50

 resourceCulture

0.48

 esempi

0.47

 الحره

0.46

 consectetur

0.46

verwijspagina

0.45

Viitteet

0.44

Autoritní

0.43

IsPostBack

0.43

Activations Density 0.067%

references to notable figures and scholars in various disciplines

names of people

No Comments

No Known Activations