INDEX

Explanations

attends to tokens related to mathematical calculations or operations from tokens that are further defined later in the sequence

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

0:0.06

1:0.08

2:0.07

3:0.11

4:0.11

5:0.04

6:0.34

7:0.16

Negative Logits

 resourceCulture

-0.47

InjectAttribute

-0.45

ResumeLayout

-0.44

 ProtoMessage

-0.44

 ***!

-0.40

rrggbb

-0.39

yntaxException

-0.38

UnusedPrivate

-0.38

 MonoBehaviour

-0.37

WriteBarrier

-0.36

POSITIVE LOGITS

 dimaksud

0.28

?}

0.28

 teto

0.28

conceito

0.28

 hoofd

0.28

 لیے

0.28

 Waray

0.28

 Torrey

0.27

dı

0.27

عة

0.27

Activations Density 0.597%

attends to tokens related to mathematical calculations or operations from tokens that are further defined later in the sequence

No Comments

No Known Activations