INDEX

Explanations

references to specific articles or discussions with notable significance

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

question or how phrases

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

richTextPanel

-0.42

 <<<<<<<<<<<<<<

-0.41

 상세

-0.41

\{\\

-0.40

Personendaten

-0.40

țion

-0.39

 tilgjenge

-0.36

麽

-0.35

 explicit

-0.34

们

-0.34

POSITIVE LOGITS

ItemBackground

0.60

 Italijanski

0.60

 незавершена

0.57

 Comprometido

0.48

@[+][

0.46

rootDir

0.46

OGND

0.44

aarrggbb

0.42

 EconPapers

0.42

 Италијани

0.41

Activations Density 0.008%

references to specific articles or discussions with notable significance

question or how phrases

No Comments

No Known Activations