INDEX

Explanations

attends to numerical values associated with ingredients and measures from each itemized list or category entry

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.33

1:0.23

2:0.08

3:0.06

4:0.06

5:0.06

6:0.05

7:0.10

Negative Logits

 […]

-0.26

PackageInfo

-0.26

 TextAppearance

-0.25

amazonaws

-0.24

🏻‍♀️

-0.24

✨:

-0.24

,...

-0.23

 onAnimation

-0.23

 ['./

-0.23

-...

-0.22

POSITIVE LOGITS

 متعلقه

0.39

Hauptartikel

0.38

rungsseite

0.38

 تانيه

0.38

ußer

0.34

 AssemblyTitle

0.32

lossians

0.32

 Aggi

0.32

ỡng

0.31

TypedDataSet

0.31

Activations Density 0.549%

attends to numerical values associated with ingredients and measures from each itemized list or category entry

No Comments

No Known Activations