INDEX

Explanations

attends to tokens that are function keywords or related to programming structures from subsequent tokens that are typically parts of expressions or function calls

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

0:0.16

1:0.14

2:0.15

3:0.05

4:0.13

5:0.21

6:0.04

7:0.07

Negative Logits

Демографія

-0.22

EXTERN

-0.22

millimeters

-0.21

 BorderRadius

-0.20

upuncture

-0.20

衛門

-0.20

 jspb

-0.20

 estre

-0.20

 anyone

-0.19

]-->

-0.19

POSITIVE LOGITS

!*\

0.32

Viitteet

0.29

ControllerAdvice

0.28

giz

0.28

 resourceCulture

0.28

adpleegd

0.28

клопе

0.28

 Meksiku

0.28

 Roskov

0.27

 Malt

0.27

Activations Density 0.037%

attends to tokens that are function keywords or related to programming structures from subsequent tokens that are typically parts of expressions or function calls

No Comments

No Known Activations