INDEX

Explanations

references to architectural or legal concepts and actions

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

legal and conditional terms

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

 autorytatywna

-0.63

 pinulongan

-0.54

 AssemblyCulture

-0.52

 للاسماء

-0.50

czegóły

-0.46

出版年

-0.45

IsMutable

-0.45

umumkan

-0.44

astify

-0.43

webElementXpaths

-0.43

POSITIVE LOGITS

Bav

0.42

alarm

0.42

rainy

0.40

 prospec

0.40

potranspiration

0.40

 preseason

0.39

Laughs

0.39

NewLine

0.39

 line

0.39

dominant

0.39

Activations Density 0.126%

references to architectural or legal concepts and actions

legal and conditional terms

No Comments

No Known Activations