INDEX

Explanations

function-related keywords and commands in coding contexts

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

+#+

-0.50

новниш

-0.49

cy

-0.47

 Salter

-0.47

Tembelea

-0.46

argas

-0.44

 Burnham

-0.44

ilstein

-0.44

 tumb

-0.43

menin

-0.43

POSITIVE LOGITS

 propOrder

0.62

parsedMessage

0.52

CloseOperation

0.50

LabelTagHelper

0.50

XmlAccessorType

0.50

SharedCtor

0.48

bezpiec

0.44

fromnode

0.43

 conmigo

0.41

UnusedPrivate

0.40

Activations Density 0.176%

function-related keywords and commands in coding contexts

No Comments

No Known Activations