INDEX

Explanations

repetitive phrases or terms, particularly those relating to conditions and forms

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

form or fashion

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 31-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_31/width_131k/average_l0_109

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.31.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

pondre

-0.40

cées

-0.37

 ویکی‌آمباردا

-0.35

Intérieur

-0.34

 Pacífico

-0.33

OuterClass

-0.32

 Relic

-0.31

 gelen

-0.31

 viktig

-0.31

ESSA

-0.31

POSITIVE LOGITS

脚注の使い方

0.76

 initComponents

0.52

存于互联网档案馆

0.50

__(/*!

0.48

lusconi

0.47

principalTable

0.45

 means

0.44

⇑

0.44

Bibliograf

0.44

AnchorStyles

0.44

Activations Density 0.004%

repetitive phrases or terms, particularly those relating to conditions and forms

form or fashion

No Comments

No Known Activations