INDEX

Explanations

attends to tokens that contain figures or references from subsequent 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

0:0.10

1:0.12

2:0.14

3:0.05

4:0.07

5:0.07

6:0.09

7:0.31

Negative Logits

DispatchToProps

-0.29

roduction

-0.28

createComponent

-0.27

ictionary

-0.27

ordinaire

-0.26

енча

-0.26

çu

-0.26

 viewType

-0.25

fandom

-0.25

Aza

-0.25

POSITIVE LOGITS

 فريبيس

0.31

abetes

0.31

hetics

0.30

ccoli

0.29

 otomatig

0.29

0.28

 المعيارى

0.28

InjectAttribute

0.28

المشاركات

0.27

ģ

0.27

Activations Density 0.015%

attends to tokens that contain figures or references from subsequent tokens

No Comments

No Known Activations