INDEX

Explanations

attends to tokens that indicate places or times from tokens that describe actions or identities

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

1:0.18

2:0.16

3:0.10

4:0.08

5:0.03

6:0.07

7:0.22

Negative Logits



-0.48

basicConfig

-0.38

 betweenstory

-0.36

 HomeComponent

-0.33

 виправивши

-0.32

 GenerationType

-0.32

ſelves

-0.31

 houſe

-0.31

 Reſ

-0.30

endphp

-0.30

POSITIVE LOGITS

elis

0.25

stateMutability

0.25

 allmän

0.24

 gmina

0.23

͘

0.23

udur

0.23

vergne

0.23

 alder

0.23

 comércio

0.23

än

0.23

Activations Density 0.099%

attends to tokens that indicate places or times from tokens that describe actions or identities

No Comments

No Known Activations