INDEX

Explanations

phrases that indicate high ranking or top-quality elements

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

top rated or best

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 ſta

-0.44

webElementXpaths

-0.43

 therefrom

-0.43

 Anſ

-0.42

تقاوى

-0.42

ьаж

-0.42

EReference

-0.42

Questions

-0.42

 Straß

-0.41

 potest

-0.41

POSITIVE LOGITS

 للاسماء

0.56

 فريبيس

0.54

 AttributeSet

0.46

 שוליים

0.44

DockStyle

0.42

 NSCoder

0.41

匿名使用者

0.39

Tembelea

0.38

 ligiloj

0.38

OOTDTY

0.36

Activations Density 0.000%

phrases that indicate high ranking or top-quality elements

top rated or best

No Comments

No Known Activations