INDEX

Explanations

specific dates and references to the Bootstrap framework

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

versions and urls

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

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

er

-0.78

-0.61

ER

-0.53

 tongs

-0.52

ed

-0.52

mybatisplus

-0.49

house

-0.49

wy

-0.45

ヤー

-0.45

󠁿

-0.44

POSITIVE LOGITS

 autorytatywna

0.77

 ویکی‌پدی

0.75

tréal

0.72

viewWillAppear

0.72

SequentialGroup

0.70

 ivelany

0.68

ecake

0.68

 ostavi

0.67

 propOrder

0.67

 NgModule

0.67

Activations Density 0.001%

specific dates and references to the Bootstrap framework

versions and urls

No Comments

No Known Activations