INDEX

Explanations

patterns associated with specific coding constructs or syntax elements

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

periods after abbreviations and initial letters

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 alambre

-0.56

RegressionTest

-0.54

 itſelf

-0.52

 montón

-0.49

Amical

-0.48

 themſelves

-0.48

 pasaje

-0.47

ScopeManager

-0.47

 exceso

-0.46

 mundiales

-0.46

POSITIVE LOGITS

jsxFileName

0.51

########.

0.46

transQ

0.44

setViewportView

0.41

béco

0.39

XmlAccessType

0.37

 الرياضيه

0.36

oredCriteria

0.36

ByUrl

0.35

WithIOException

0.35

Activations Density 0.303%

patterns associated with specific coding constructs or syntax elements

periods after abbreviations and initial letters

No Comments

No Known Activations

patterns associated with specific coding constructs or syntax elements

periods after abbreviations and initial letters

No Comments

No Known Activations