INDEX

Explanations

words related to research methodologies and findings

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

association with

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

 mostrarán

-0.34

 protoimpl

-0.33

overse

-0.29

Major

-0.29

 Longman

-0.28

 phare

-0.28

 Spisak

-0.27

 endregion

-0.27

évaluateur

-0.27

zbęd

-0.27

POSITIVE LOGITS

 InputDecoration

0.64

ViewFeatures

0.58

 дописавши

0.57

 مشين

0.50

DockStyle

0.50

 gyhoeddwyd

0.49

GIH

0.48

 незавершена

0.47

orianCalendar

0.46

 getItemId

0.46

Activations Density 0.051%

words related to research methodologies and findings

association with

No Comments

No Known Activations