INDEX

Explanations

attends to function-related tokens from adjacent programming or timing-related 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

0:0.12

1:0.17

2:0.14

3:0.09

4:0.14

5:0.08

6:0.09

7:0.13

Negative Logits

EndGlobalSection

-0.43

 ISNI

-0.35

cadilly

-0.35

AddTagHelper

-0.33

 "..\..\..\

-0.30

contentLoaded

-0.29

shadowRadius

-0.28

новниш

-0.28

ArgsConstructor

-0.27

RotationOrder

-0.27

POSITIVE LOGITS

やかに

0.26

schema

0.24

“

0.24

son

0.24

 đồ

0.24

ımda

0.23

ist

0.22

kunt

0.22

 plomb

0.22

е

0.22

Activations Density 0.036%

attends to function-related tokens from adjacent programming or timing-related tokens

No Comments

No Known Activations