INDEX

Explanations

attends to tokens associated with possession from related tokens indicating an entity or item

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

0:0.12

1:0.15

2:0.10

3:0.05

4:0.04

5:0.02

6:0.05

7:0.43

Negative Logits

PYX

-0.34

vician

-0.34

dders

-0.32

 récupération

-0.32

ccoli

-0.32

 Lehre

-0.32

ARGS

-0.32

 cauza

-0.32

 bäst

-0.31

glBindBuffer

-0.31

POSITIVE LOGITS

 مشين

0.26

 GenerationType

0.25

RatingBar

0.25

 AppCompatTheme

0.24

WebControls

0.24

 LEYENDO

0.24

횟

0.23

atimes

0.23

 MonoBehaviour

0.23

onder

0.23

Activations Density 0.690%

attends to tokens associated with possession from related tokens indicating an entity or item

No Comments

No Known Activations