INDEX

Explanations

key terms and phrases related to important findings or contributions in scientific work

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

purpose of new schemes

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

fjspx

-0.44

 ویکی‌پدیا

-0.42

原始内容存档于

-0.36

山

-0.36

 deswegen

-0.35

 Италијани

-0.35

цездатний

-0.35

amateur

-0.35

AppMethodBeat

-0.34

èdia

-0.33

POSITIVE LOGITS

styleType

0.53

andExpect

0.50

govine

0.44

 sánh

0.44

EndProject

0.44

 TestBed

0.43

 BoxDecoration

0.42

enterOuterAlt

0.41

<?

0.41

Халык

0.41

Activations Density 0.006%

key terms and phrases related to important findings or contributions in scientific work

purpose of new schemes

No Comments

No Known Activations