INDEX

Explanations

attends to numeric tokens from corresponding items listed in sequences

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

0:0.13

1:0.13

2:0.11

3:0.09

4:0.22

5:0.14

6:0.07

7:0.07

Negative Logits

loroethene

-0.34

 AssemblyTitle

-0.31

 CURIAM

-0.29

)))));

-0.29

ugier

-0.28



-0.28

 Conroy

-0.27

MENAFN

-0.26

atherapy

-0.26

nloa

-0.26

POSITIVE LOGITS

@"/

0.38

UnitTesting

0.37

ReusableCell

0.37

 HasFactory

0.36

Autoritní

0.34

 VIAF

0.34

Xna

0.33

Revenir

0.32

 EconPapers

0.32

gnügen

0.31

Activations Density 0.897%

attends to numeric tokens from corresponding items listed in sequences

No Comments

No Known Activations