INDEX

Explanations

attends to a variety of action-related tokens from various contextual tokens, often focusing on legal or formal proceedings

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

0:0.09

1:0.10

2:0.13

3:0.07

4:0.03

5:0.01

6:0.06

7:0.49

Negative Logits

MigrationBuilder

-0.47

webElementXpaths

-0.41

 Landis

-0.35

obel

-0.34

InjectAttribute

-0.33

ItemBackground

-0.33

 transfieras

-0.32

SpringBootTest

-0.32

lapsingToolbar

-0.31

IsMutable

-0.31

POSITIVE LOGITS

SequentialGroup

0.32

 resourceCulture

0.30

iyle

0.30

mogorov

0.29

 initComponents

0.28

 Accesat

0.27

berdayakan

0.27

}{*}{

0.26

writeFieldEnd

0.26

клопе

0.26

Activations Density 0.843%

attends to a variety of action-related tokens from various contextual tokens, often focusing on legal or formal proceedings

No Comments

No Known Activations