INDEX

Explanations

attends to a specific keyword or button expression from general output text

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

1:0.10

2:0.12

3:0.05

4:0.06

5:0.03

6:0.10

7:0.43

Negative Logits

 autorytatywna

-0.35

SBATCH

-0.32

DataAnnotations

-0.31

tagHelperRunner

-0.31

manteau

-0.28

 CreateTagHelper

-0.28

getColumnIndex

-0.28

ņi

-0.27

 vroe

-0.26

 Normdatei

-0.26

POSITIVE LOGITS

 NSCoder

0.41

JsonKey

0.35

lise

0.28

uesia

0.28

geord

0.28

^{[

0.28

ніципалі

0.27

 semej

0.26

 Spoilers

0.26

uhr

0.26

Activations Density 0.166%

attends to a specific keyword or button expression from general output text

No Comments

No Known Activations