INDEX

Explanations

attends to tokens marked with specific semantic meaning related to objects from subsequent tokens referencing contextually relevant items

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

0:0.11

1:0.15

2:0.16

3:0.09

4:0.12

5:0.03

6:0.10

7:0.19

Negative Logits

tagHelperRunner

-0.32

 تضيفلها

-0.28

MessageTagHelper

-0.28

PageRoute

-0.27

 خارجية

-0.27

AndWait

-0.27

parsedMessage

-0.26

 calendriers

-0.26

ClassLoader

-0.26

GOTREF

-0.26

POSITIVE LOGITS

rscheinlich

0.31

 Doherty

0.28

 betweenstory

0.28

ArgsConstructor

0.28

Bios

0.27

AOC

0.27

 bParam

0.27

 Lons

0.26

 Bios

0.26

ientôt

0.26

Activations Density 0.252%

attends to tokens marked with specific semantic meaning related to objects from subsequent tokens referencing contextually relevant items

No Comments

No Known Activations