INDEX

Explanations

attends to various scientific terms or elements from corresponding numerical references or related 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

0:0.12

1:0.16

2:0.13

3:0.09

4:0.10

5:0.02

6:0.11

7:0.22

Negative Logits

TagMode

-0.26

 <=",

-0.26

elemField

-0.26

 lenker

-0.26

 AssemblyTitle

-0.25

lihatan

-0.25

@"/

-0.25

ьаж

-0.25

<!--[

-0.24

Jeografia

-0.24

POSITIVE LOGITS

hamshire

0.35

mew

0.31

ByteBuf

0.31

ViewFeatures

0.31

esserung

0.30

doria

0.30

APORE

0.30

 initComponents

0.29

 phylog

0.28

Batis

0.28

Activations Density 0.903%

attends to various scientific terms or elements from corresponding numerical references or related tokens

No Comments

No Known Activations