INDEX

Explanations

attends to programming-related tokens from mathematical or technical 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

0:0.08

1:0.12

2:0.17

3:0.08

4:0.09

5:0.03

6:0.18

7:0.20

Negative Logits

 '.';

-0.35

.[/

-0.35

CELLANEOUS

-0.32

engkapi

-0.31

PROVIDER

-0.29

')['

-0.29

RegressionTest

-0.29

ləş

-0.29

FFIX

-0.28

--]

-0.28

POSITIVE LOGITS

 فريبيس

0.39

 Infórmanos

0.30

QtWidgets

0.29

Bor

0.29

tagHelperRunner

0.28

ărilor

0.27

 onlyOwner

0.26

atorio

0.26

BAB

0.26

hint

0.26

Activations Density 0.001%

attends to programming-related tokens from mathematical or technical tokens

No Comments

No Known Activations