INDEX

Explanations

code-related terms and functions within a programming context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 17-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_17/width_16k/average_l0_73

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.17.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

 “…

-0.35

 defaultstate

-0.35

⅕

-0.34

 anstatt

-0.32

 Chwiliwch

-0.31

¾

-0.30

OptionPane

-0.30

⅗

-0.29

 finger

-0.29

 الض

-0.29

POSITIVE LOGITS

 propOrder

0.85

XmlAccessorType

0.73

 surla

0.65

 &___

0.64

LookAnd

0.64

IndentedString

0.64

 Италијани

0.59

OrBuilder

0.59

 AssemblyCulture

0.59

 lenker

0.58

Activations Density 0.350%

code-related terms and functions within a programming context

No Comments

No Known Activations