INDEX

Explanations

instances of the term "number" and its various forms associated with numerical quantities

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.20.hook_resid_post

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

Negative Logits

 Belajar

-0.45

 Conroy

-0.44

Früh

-0.43

 Driscoll

-0.43

aprend

-0.41

 glean

-0.41

 tepi

-0.40

 изучения

-0.40

exp

-0.40

POSITIVE LOGITS

 Number

1.30

 number

1.28

 getNumber

1.24

Number

1.23

 NUMBER

1.20

NUMBER

1.14

 Numbers

1.11

getNumber

1.10

 numbers

1.08

number

1.07

Activations Density 0.179%

instances of the term "number" and its various forms associated with numerical quantities

No Comments

No Known Activations