INDEX

Explanations

attends to tokens related to names or titles that are associated with the token "of" from later tokens containing actions or descriptions

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

0:0.13

1:0.14

2:0.36

3:0.04

4:0.04

5:0.02

6:0.04

7:0.18

Negative Logits

 tartalomajánló

-0.36

ethene

-0.33

inergic

-0.28

 kasarigan

-0.28

});

-0.28

HCR

-0.27

 amorti

-0.26

rovnik

-0.26

AMT

-0.25

 anthrop

-0.25

POSITIVE LOGITS

modelBuilder

0.44

 which

0.35

gink

0.34

 arşivlendi

0.33

EMPL

0.31

 lead

0.30

interopRequire

0.30

blockList

0.30

lead

0.29

MessageOf

0.29

Activations Density 0.188%

attends to tokens related to names or titles that are associated with the token "of" from later tokens containing actions or descriptions

No Comments

No Known Activations