INDEX

Explanations

references to prior knowledge or arrangements related to planning and preparation

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

planning beforehand

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

pæ

-0.40

TagMode

-0.39

mã

-0.39

kháu

-0.38

zzleHttp

-0.38

 PeEnEo

-0.37

 Kebijakan

-0.37

bien

-0.37

Datuak

-0.36

ungere

-0.36

POSITIVE LOGITS

.*")]

0.45

 beforehand

0.45

ViewFeatures

0.41

+:+

0.40

ByEmail

0.40

ScopeManager

0.40

JpaRepository

0.39

 booking

0.39

 Roskov

0.39

 дописавши

0.39

Activations Density 0.038%

references to prior knowledge or arrangements related to planning and preparation

planning beforehand

No Comments

No Known Activations