INDEX

Explanations

Java programming constructs and control flow statements

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

function return statements

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

Становништво

-0.65

+#+#

-0.63

 ujednoznacz

-0.57

 EconPapers

-0.57

 الرياضيه

-0.56

 Roskov

-0.56

ThroughAttribute

-0.55

 становника

-0.55

elemField

-0.54

Personensuche

-0.52

POSITIVE LOGITS

 unnecessary

0.35

 Hence

0.34

 utafitiHapana

0.34

<?

0.33

 After

0.32

 Lastly

0.32

 hence

0.32

After

0.32

Attra

0.32

 Maybe

0.31

Activations Density 0.050%

Java programming constructs and control flow statements

function return statements

No Comments

No Known Activations