INDEX

Explanations

attends to specific code-related tokens from associated syntactical or data structure 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

2:0.11

3:0.20

4:0.08

5:0.07

6:0.17

7:0.17

Negative Logits

ρίς

-0.26

gridx

-0.26

enko

-0.23

 saper

-0.23

 acides

-0.22

 متعلقه

-0.22

쳐

-0.22

er

-0.22

 ratón

-0.22

 Lehrer

-0.22

POSITIVE LOGITS

 XNUMX

0.40

tableFuture

0.39

mobileqq

0.37

 numberWith

0.36

NUMX

0.35

 FetchType

0.35

 ujednoznacz

0.34

 يتيمه

0.33

hidupan

0.32

TagMode

0.32

Activations Density 0.000%

attends to specific code-related tokens from associated syntactical or data structure tokens

No Comments

No Known Activations