INDEX

Explanations

attends to instances of "positive" from contexts involving evaluation or detection methods

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

0:0.08

1:0.08

2:0.09

3:0.10

4:0.07

5:0.02

6:0.19

7:0.32

Negative Logits

 enää

-0.28

化为

-0.27

 Ababa

-0.27

fbox

-0.26

SerializedName

-0.25

fiche

-0.25

GenerationType

-0.25

 baratas

-0.24

 Hinton

-0.24

bindung

-0.24

POSITIVE LOGITS

 resourceCulture

0.40

IntoConstraints

0.33

 referenties

0.33

 Wiktionnaire

0.31

twimg

0.30

 الدولى

0.29

NameInMap

0.29

illust

0.28

 CreateTagHelper

0.28

Населення

0.28

Activations Density 1.579%

attends to instances of "positive" from contexts involving evaluation or detection methods

No Comments

No Known Activations