INDEX

Explanations

attends to mathematical tokens or equations from tokens that refer to their positional or relational context

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

0:0.08

1:0.11

2:0.08

3:0.11

4:0.10

5:0.05

6:0.28

7:0.15

Negative Logits

…~

-0.41

%%

-0.35

rha

-0.35

crans

-0.35

verwijspagina

-0.34

 inútil

-0.34

everywhere

-0.34

úblico

-0.34

Groetjes

-0.33

rativo

-0.33

POSITIVE LOGITS

 متعلقه

0.41

ReusableCell

0.39

Geplaatst

0.38

Xna

0.38

AutoScaleMode

0.36

LookAnd

0.35

tagHelperRunner

0.35

AspNetCore

0.34

adon

0.33

ineno

0.33

Activations Density 0.022%

attends to mathematical tokens or equations from tokens that refer to their positional or relational context

No Comments

No Known Activations