INDEX

Explanations

attends to the last token from the preceding sentence from earlier 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

0:0.07

1:0.10

2:0.08

3:0.11

4:0.18

5:0.08

6:0.24

7:0.10

Negative Logits

cèse

-0.37

 Italijanski

-0.35

 intersti

-0.35

NameInMap

-0.35

TintMode

-0.34

Vidite

-0.34

يكب

-0.34

 Мексичка

-0.34

BufferException

-0.33

 springfox

-0.32

POSITIVE LOGITS

 />";

0.30

 sauter

0.30

 électro

0.28

 volontaire

0.26

">//

0.25

läng

0.25

ורת

0.25

 råd

0.25

 odkazy

0.25

 béné

0.24

Activations Density 0.245%

attends to the last token from the preceding sentence from earlier tokens

No Comments

No Known Activations