INDEX

Explanations

attends to the token representing a suffix from tokens that are emphasized in a positive context

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

0:0.11

1:0.13

2:0.11

3:0.12

4:0.12

5:0.11

6:0.11

7:0.15

Negative Logits

Sucesor

-0.36

expandindo

-0.36

تقاوى

-0.34

protoimpl

-0.33

 بتاريخ

-0.33

textAppearance

-0.32

########.

-0.31

ⓧ

-0.29

MERCIAL

-0.28

minton

-0.28

POSITIVE LOGITS

 بيها

0.29

 للمعارف

0.27

 كومونز

0.27

gehe

0.24

 Hare

0.24

">//

0.24

 uwagę

0.23

/**

0.23

CWE

0.23

 inov

0.23

Activations Density 0.026%

attends to the token representing a suffix from tokens that are emphasized in a positive context

No Comments

No Known Activations