INDEX

Explanations

attends to specific code or directive tokens from preceding contextual or structural markers

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

0:0.19

1:0.10

2:0.34

3:0.08

4:0.07

5:0.03

6:0.06

7:0.10

Negative Logits

 "..\..\

-0.32

帖最后由

-0.31

+:+

-0.29

Хьажоргаш

-0.29

queta

-0.27

最快更新

-0.27

@"/

-0.27

saraba

-0.26

 TextAppearance

-0.26

 also

-0.26

POSITIVE LOGITS

 يتيمه

0.38

WebVitals

0.37

XtraEditors

0.33

 '\\;'

0.33

 Theſe

0.30

ppins

0.27

VIAF

0.26

IActionResult

0.25

 snippetHide

0.25

 Schlu

0.25

Activations Density 0.007%

attends to specific code or directive tokens from preceding contextual or structural markers

No Comments

No Known Activations