INDEX

Explanations

attends to tokens indicating peak or critical values from tokens suggesting subsequent limits or related effects

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

0:0.27

1:0.22

2:0.17

3:0.09

4:0.06

5:0.03

6:0.04

7:0.08

Negative Logits

AnchorStyles

-0.38

ConstraintMaker

-0.35

 snippetHide

-0.35

 ExecuteAsync

-0.34

 consultato

-0.33

awtextra

-0.30

EndInit

-0.29

ViewById

-0.29

 оригіналу

-0.29

 становника

-0.28

POSITIVE LOGITS

 sahiptir

0.21

Aspiration

0.20

 cser

0.20

 trebui

0.20

 sfor

0.20

cipar

0.19

 Wassers

0.19

 Städ

0.19

ArgumentParser

0.19

 szív

0.18

Activations Density 0.595%

attends to tokens indicating peak or critical values from tokens suggesting subsequent limits or related effects

No Comments

No Known Activations