INDEX

Explanations

attends to gluing tokens from various earlier and later tokens in sequences

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

0:0.13

1:0.11

2:0.47

3:0.06

4:0.05

5:0.02

6:0.03

7:0.08

Negative Logits

InstrumentedTest

-0.43

 bezeichneter

-0.40

CloseOperation

-0.40

دانشنامهٔ

-0.39

WindowConstants

-0.37

 للاسماء

-0.37

 समीक्षाएं

-0.33

يميديا

-0.32

numerusform

-0.32

Hentet

-0.32

POSITIVE LOGITS

ndose

0.28

 écout

0.25

 <=",

0.24

TabView

0.24

روج

0.23

ستی

0.23

RTEX

0.23

}){

0.23

 []),

0.23

Gree

0.22

Activations Density 0.801%

attends to gluing tokens from various earlier and later tokens in sequences

No Comments

No Known Activations