INDEX

Explanations

attends to complex mathematical symbols or terms from simplified or other contributing mathematical variables

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

0:0.17

1:0.19

2:0.13

3:0.07

4:0.08

5:0.03

6:0.09

7:0.20

Negative Logits

parsedMessage

-0.40

ể

-0.36

 Balo

-0.35

forth

-0.35

Errorf

-0.34

 Архів

-0.34

存于互联网档案馆

-0.32

ölkerung

-0.32

 tand

-0.31

lut

-0.31

POSITIVE LOGITS

AndEndTag

0.31

눠

0.30

Grüsse

0.28

 صوتيه

0.27

DoubleQuotes

0.27

 kasarigan

0.26

akujem

0.26

WireFormatLite

0.26

SharedDtor

0.26

mochila

0.25

Activations Density 2.151%

attends to complex mathematical symbols or terms from simplified or other contributing mathematical variables

No Comments

No Known Activations