INDEX

Explanations

attends to specific tokens marked with double characters from subsequent tokens containing a single character

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.07

1:0.06

2:0.07

3:0.04

4:0.03

5:0.02

6:0.08

7:0.60

Negative Logits

 الحره

-0.31

 препратки

-0.29

WithIdentifier

-0.28

 tartalomajánló

-0.27

مرح

-0.26

 Lingkungan

-0.26

 promozione

-0.26

Encyklopedia

-0.25

ásico

-0.25

 EXPERIMENT

-0.25

POSITIVE LOGITS

Vidite

0.37

 Wikimedijinoj

0.33

brainly

0.32

abestanden

0.28

();)

0.28

 preocup

0.27

}*/

0.27

aurait

0.26

pixabay

0.26

 Italijani

0.25

Activations Density 0.583%

attends to specific tokens marked with double characters from subsequent tokens containing a single character

No Comments

No Known Activations