INDEX

Explanations

punctuation marks, particularly commas

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

family members and spouses

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

 ویکی‌پدی

-0.82

 ProtoMessage

-0.64

Personendaten

-0.64

脚注の使い方

-0.61

 pinulongan

-0.60

oa̍t

-0.59

DllImport

-0.54

 Italijanski

-0.54

XmlAccessType

-0.52

AccessorTable

-0.52

POSITIVE LOGITS

gegevens

0.41

 whose

0.41

="#"

0.41

 consuls

0.40

who

0.40

unzel

0.40

riwal

0.39

 Argentino

0.38

whose

0.38

 Sosial

0.37

Activations Density 0.014%

punctuation marks, particularly commas

family members and spouses

No Comments

No Known Activations