INDEX

Explanations

references to user instructions and advice

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

advice, instructions, and recommendations

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

language that signals instructional or advisory content—requests for help, guidance, or information and references to documentation, examples, blogs, or advice.

oai_token-act-pair · gpt-5 Triggered by @vetterc0

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

kwds

-0.43

 calculées

-0.37

 managed

-0.29

レーキ

-0.28

 вече

-0.28

},[])

-0.27

"",

-0.27

仁

-0.27

okkal

-0.27

 meng

-0.26

POSITIVE LOGITS

 advices

0.70

 advice

0.68

featureID

0.67

 recommendations

0.61

advice

0.60

Consejos

0.59

 ADVICE

0.59

recommendations

0.59

HtmlAttribute

0.59

instructions

0.57

Activations Density 0.182%

references to user instructions and advice

advice, instructions, and recommendations

language that signals instructional or advisory content—requests for help, guidance, or information and references to documentation, examples, blogs, or advice.

No Comments

No Known Activations