INDEX

Explanations

attends to the token "here" from tokens that provide references or additional information

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

0:0.15

1:0.18

2:0.34

3:0.07

4:0.05

5:0.02

6:0.05

7:0.11

Negative Logits

kaka

-0.28

ggior

-0.28

UrlResolution

-0.28

 templ

-0.28

зульта

-0.27

nocześnie

-0.27

__(/*!

-0.27

EREF

-0.27

usepackage

-0.27

 gamba

-0.27

POSITIVE LOGITS

 uska

0.36

prefixer

0.29

AndEndTag

0.29

RelativeTo

0.27

lapsingToolbar

0.25

 cardíaca

0.25

Véxase

0.24

("-");

0.24

 surla

0.24

 useDispatch

0.24

Activations Density 0.232%

attends to the token "here" from tokens that provide references or additional information

No Comments

No Known Activations