INDEX

Explanations

punctuation marks, particularly commas

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

start of turn user

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

 éste

-0.36

 fondos

-0.35

AutoModerator

-0.31

 fallecimiento

-0.31

 BETTER

-0.30

 altogether

-0.30

InstrumentedTest

-0.29

 rebajas

-0.29

 celui

-0.29

FAV

-0.28

POSITIVE LOGITS

 للاسماء

0.71

بوابة

0.66

Portail

0.65

tagPool

0.64

AnchorStyles

0.63

fjspx

0.63

$_(

0.63

 JpaRepository

0.61

 NgModule

0.60

NgModule

0.60

Activations Density 0.000%

punctuation marks, particularly commas

start of turn user

No Comments

No Known Activations