INDEX

Explanations

attends to pairs of tokens that are both mathematical or technical-related, focusing on those with corresponding punctuation marks

oai_attention-head · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-att-16k

Configuration

google/gemma-scope-2b-pt-att/layer_0/width_16k/average_l0_104

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.0.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

0:0.06

1:0.61

2:0.03

3:0.02

4:0.08

5:0.07

6:0.02

7:0.07

Negative Logits

 Paglinawan

-0.60

 Réponses

-0.58

queryInterface

-0.56

 ModelExpression

-0.55

ьаж

-0.54

Билгалдахарш

-0.54

SharedDtor

-0.53

 propOrder

-0.53

 ProtoMessage

-0.53

POSITIVE LOGITS

 ſhe

0.43

co

0.42

 himſelf

0.41

 inſ

0.39

 purpoſe

0.39

 themſelves

0.38

 Diſ

0.37

 Inſ

0.37

 Anſ

0.37

cl

0.36

Activations Density 0.040%

attends to pairs of tokens that are both mathematical or technical-related, focusing on those with corresponding punctuation marks

No Comments

No Known Activations