INDEX

Explanations

attends to tokens that represent document types or formats from tokens that denote specific references or identifiers

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

0:0.10

1:0.12

2:0.06

3:0.07

4:0.07

5:0.10

6:0.31

7:0.14

Negative Logits

IntoConstraints

-0.38

GEBURTSDATUM

-0.37

glGen

-0.35

 Normdatei

-0.34

EndContext

-0.32

aarrggbb

-0.32

gorithm

-0.32

stanbul

-0.31

 bloco

-0.31

 stride

-0.31

POSITIVE LOGITS

antd

0.32

ंदीखरीदारी

0.31

findpost

0.30

期刊论文

0.29

inden

0.28

/*

0.28

setDo

0.27

tagHelperRunner

0.27

tvguidetime

0.26

UserScript

0.26

Activations Density 0.001%

attends to tokens that represent document types or formats from tokens that denote specific references or identifiers

No Comments

No Known Activations