INDEX

Explanations

attends to the token "last" from the subsequent token "first."

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.31

1:0.22

2:0.15

3:0.07

4:0.06

5:0.03

6:0.03

7:0.09

Negative Logits

AndEndTag

-0.36

IntoConstraints

-0.35

aarrggbb

-0.35

OGND

-0.34

">//

-0.34

omock

-0.33

 ویکی‌پدیا

-0.31

 beginnetje

-0.30

featureID

-0.30

CPtr

-0.30

POSITIVE LOGITS

σιμοποι

0.26

 digress

0.24

 berarti

0.24

følgelig

0.24

 sababu

0.23

ñora

0.23

 widać

0.22

íš

0.22

queiro

0.22

ğından

0.22

Activations Density 0.224%

attends to the token "last" from the subsequent token "first."

No Comments

No Known Activations