INDEX

Explanations

root words and their morphological variants related to specific actions or characteristics

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

calculated, stabilizer, critical, containing, recommend, disciplined

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

 propOrder

-0.58

GEBURTSDATUM

-0.54

 arşivlendi

-0.51

currentColor

-0.47

ModelSerializer

-0.46

Portail

-0.46

таратура

-0.44

MLLoader

-0.44

CPtr

-0.44

তথ্যসূত্র

-0.44

POSITIVE LOGITS

BibitemShut

0.52

kespea

0.51

printStats

0.50

citenamefont

0.49

bibfield

0.49

bibinfo

0.49

ihnach

0.48

eronau

0.47

cknow

0.47

bibnamefont

0.47

Activations Density 0.059%

root words and their morphological variants related to specific actions or characteristics

calculated, stabilizer, critical, containing, recommend, disciplined

No Comments

No Known Activations