INDEX

Explanations

attends to action-related tokens from specific operational or conceptual terms

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

0:0.12

1:0.11

2:0.16

3:0.06

4:0.04

5:0.02

6:0.07

7:0.38

Negative Logits

toMatchSnapshot

-0.35

ThroughAttribute

-0.34

)];

-0.33

 pleaſure

-0.30

!);

-0.30

 Reſ

-0.29

AndEndTag

-0.29

FormTagHelper

-0.29

'));

-0.28

 gynhyrchwyd

-0.28

POSITIVE LOGITS

forRoot

0.31

 elegans

0.29

 οπο

0.28

навли

0.27

 expédi

0.26

ggiamento

0.26

kloped

0.26

//

0.26

 Horoscope

0.26

entistas

0.25

Activations Density 0.495%

attends to action-related tokens from specific operational or conceptual terms

No Comments

No Known Activations