INDEX

Explanations

attends to specific measurements related to weight from token pairs featuring weight indicators

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

0:0.05

1:0.11

2:0.06

3:0.04

4:0.16

5:0.45

6:0.04

7:0.05

Negative Logits

LookAnd

-0.33

+#+#

-0.29

 Form

-0.26

usammen

-0.25

trum

-0.25

Ob

-0.25

нуться

-0.25

 trom

-0.24

tru

-0.24

heda

-0.24

POSITIVE LOGITS

 وتسجيلات

0.27

 HasFactory

0.26

})*/

0.25

 fédé

0.24

 Quellen

0.24

 Catania

0.24

 CUCET

0.23

 inspiração

0.23

Дереккөздер

0.23

featureID

0.23

Activations Density 0.173%

attends to specific measurements related to weight from token pairs featuring weight indicators

No Comments

No Known Activations