INDEX

Explanations

attends to non-related terms from related terms

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

0:0.14

1:0.16

2:0.46

3:0.05

4:0.03

5:0.03

6:0.02

7:0.08

Negative Logits

 nakalista

-0.44

IsContent

-0.44

MemoryWarning

-0.43

VIAF

-0.41

SourceChecksum

-0.38

nologue

-0.37

IndentedString

-0.36

 يتيمه

-0.36

 kasarigan

-0.35

ValueStyle

-0.35

POSITIVE LOGITS

0.26

цепт

0.26

 thương

0.24

vella

0.23

 tanda

0.22

blers

0.22

 getNum

0.22

 extra

0.22

 that

0.22

zame

0.21

Activations Density 1.004%

attends to non-related terms from related terms

No Comments

No Known Activations