INDEX

Explanations

names of prominent individuals and notable entities

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

mixed-language names and technical terms

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

 gadgets

-0.42

 tricks

-0.40

 kids

-0.40

 trick

-0.39

 baby

-0.39

crazy

-0.39

verwijspagina

-0.39

 stuff

-0.39

naughty

-0.37

tvguidetime

-0.37

POSITIVE LOGITS

GEBURTS

0.66

jsxFileName

0.61

蚪

0.59

AnimationsModule

0.57

MigrationBuilder

0.56

complexType

0.55

exitRule

0.54

 TextAlign

0.54

IActionResult

0.52

XmlAccessType

0.51

Activations Density 0.156%

names of prominent individuals and notable entities

mixed-language names and technical terms

No Comments

No Known Activations