INDEX

Explanations

attends to specific terms or phrases that have defined qualities or implications from relational or contextual tokens

oai_attention-head · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 10-gemmascope-att-16k

Configuration

google/gemma-scope-2b-pt-att/layer_10/width_16k/average_l0_70

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.10.attn.hook_z

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.30

1:0.28

2:0.10

3:0.06

4:0.06

5:0.02

6:0.06

7:0.07

Negative Logits

 kasarigan

-0.30

 naturali

-0.24

+"/

-0.24

Datuak

-0.23

__':

-0.23

/*

-0.23

("/:

-0.21

entar

-0.21

 naturales

-0.20

 Fils

-0.20

POSITIVE LOGITS

 صوتيه

0.32

makeConstraints

0.29

ConstraintMaker

0.27

InstanceState

0.27

PrototypeOf

0.27

INCREF

0.25

 ویکی‌پدیای

0.25

 Chau

0.25

 JOptionPane

0.23

amaño

0.23

Activations Density 1.479%

attends to specific terms or phrases that have defined qualities or implications from relational or contextual tokens

No Comments

No Known Activations