INDEX

Explanations

numeric identifiers related to parts or components within a procedural context

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

technical terms and citations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

󠁴

-0.71

/**

-0.68

IntoConstraints

-0.67

bootstrapcdn

-0.66

-0.65

PreferredItem

-0.63

/*

-0.61

✭✭

-0.58

 Geſch

-0.58

PyExc

-0.57

POSITIVE LOGITS

 isComment

0.34

 isNameExpr

0.33

期刊论文

0.32

說是

0.31

 děti

0.30

 département

0.30

 dependencia

0.29

fiya

0.29

 considerazione

0.28

 modelAndView

0.28

Activations Density 0.003%

numeric identifiers related to parts or components within a procedural context

technical terms and citations

No Comments

No Known Activations