INDEX

Explanations

attends to tokens starting with "V" from subsequent tokens that also feature "V"

oai_attention-head · gpt-4o-mini Triggered by @bot

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-att/layer_0/width_16k/average_l0_61

Prompts (Dashboard)

16,384 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.0.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

0:0.01

1:0.00

2:0.08

3:0.68

4:0.01

5:0.00

6:0.01

7:0.01

8:0.00

9:0.00

10:0.01

11:0.01

12:0.01

13:0.03

14:0.04

15:0.00

Negative Logits

 Paglinawan

-0.18

 orejas

-0.17

latego

-0.16

 alfombra

-0.16

Ecotoxicity

-0.16

 plumas

-0.16

 cabaña

-0.16

 arşivlendi

-0.16

})*/

-0.15

 aéreo

-0.15

POSITIVE LOGITS

⤹

0.28

ConstraintMaker

0.27

</strong>

0.27

:✨

0.27

VIAF

0.26

 &___

0.25

fvar

0.25

vivid

0.25

 Roskov

0.24

 Vivid

0.24

Activations Density 0.530%

attends to tokens starting with "V" from subsequent tokens that also feature "V"

No Comments

No Known Activations