INDEX

Explanations

attends to numerical examples from related social media tokens

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

1:0.25

2:0.11

3:0.08

4:0.17

5:0.12

6:0.06

7:0.09

Negative Logits

ież

-0.47

 elems

-0.37

toMatchSnapshot

-0.36

 poética

-0.35

 gyrus

-0.35

Referências

-0.34

Erreferentziak

-0.34

Lähteet

-0.33

RepeatedField

-0.33

 prolet

-0.32

POSITIVE LOGITS

enderror

0.33

COUVER

0.30

lastIndexOf

0.29

اعمال

0.29

 buch

0.28

IDAY

0.28

comod

0.27

CGRectMake

0.27

SetUp

0.26

ầy

0.26

Activations Density 0.041%

attends to numerical examples from related social media tokens

No Comments

No Known Activations