INDEX

Explanations

markers indicating the beginning of structured or organized content

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

instructions and queries related to exam procedures

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

preparing for exams or study

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_65k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

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

řevě

-0.49

DOCTYPE

-0.45

 القدم

-0.44

 invisible

-0.43

zehnten

-0.42

ankind

-0.42

 ospiti

-0.42

 büyü

-0.42

 ",";

-0.42

ワイイ

-0.42

POSITIVE LOGITS

 preparation

1.23

 Preparation

1.08

preparation

1.07

 prep

1.00

 préparation

0.97

 Vorbereitung

0.97

 PREPARATION

0.95

 preparing

0.95

 study

0.94

 prepare

0.94

Activations Density 0.040%

markers indicating the beginning of structured or organized content

instructions and queries related to exam procedures

preparing for exams or study

No Comments

No Known Activations