INDEX

Explanations

attends to directional tokens indicating spatial properties from unspecified tokens

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

0:0.11

1:0.13

2:0.13

3:0.12

4:0.13

5:0.04

6:0.14

7:0.16

Negative Logits

войства

-0.28

 protoimpl

-0.27

WithMany

-0.27

 ***!

-0.27

Conway

-0.26

Попис

-0.26

keyColumn

-0.24

 دیکھیے

-0.24

.)}

-0.24

adpleegd

-0.24

POSITIVE LOGITS

="#"><

0.35

TagMode

0.35

 bParam

0.30

 pierdas

0.30

 preocupes

0.29

ponses

0.28

 mapStateToProps

0.28

<%=

0.28

ugier

0.28

Quiénes

0.27

Activations Density 0.028%

attends to directional tokens indicating spatial properties from unspecified tokens

No Comments

No Known Activations