INDEX

Explanations

technical terms and components related to systems and programming

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

current state or usage

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

 Exactos

-0.50

 endblock

-0.49

PreExecute

-0.49

 kasarigan

-0.47

Tie

-0.47

recated

-0.45

 <<<<<<<<<<<<<<

-0.43

gameserver

-0.43

Tie

-0.43

 Walkover

-0.43

POSITIVE LOGITS

 defaultstate

0.49

 breakdowns

0.48

 المعيارى

0.38

iritto

0.38

 chrétien

0.38

arrière

0.37

følgelig

0.37

 giudi

0.36

 geçmiş

0.36

 Paglinawan

0.35

Activations Density 0.041%

technical terms and components related to systems and programming

current state or usage

No Comments

No Known Activations