INDEX

Explanations

attends to numerical tokens from constants or coefficients that are consistently used in formulas, particularly focusing on finding correlations and relations among them

oai_attention-head · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-att-16k

Configuration

google/gemma-scope-2b-pt-att/layer_0/width_16k/average_l0_104

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.0.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

0:0.56

1:0.05

2:0.02

3:0.02

4:0.08

5:0.03

6:0.03

7:0.19

Negative Logits

Hozzáférés

-0.26

 particulière

-0.26

 nenhuma

-0.25

 none

-0.24

 nessuno

-0.24

op

-0.24

-0.23

 initWith

-0.23

POSITIVE LOGITS

 المعيارى

0.64

 AssemblyCulture

0.58

rrggbb

0.49

setVerticalGroup

0.47

 betweenstory

0.47

 وتسجيلات

0.47

GraphicsUnit

0.42

########.

0.41

IndentedString

0.40

 صوتيه

0.39

Activations Density 0.067%

attends to numerical tokens from constants or coefficients that are consistently used in formulas, particularly focusing on finding correlations and relations among them

No Comments

No Known Activations