INDEX

Explanations

large sections of whitespace or empty content

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

identifiers and foreign words

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

dellín

-0.49

TypedDataSet

-0.47

 utafitiHapana

-0.47

 betweenstory

-0.45

 crucified

-0.44

FXML

-0.43

 llorando

-0.43

onViewCreated

-0.43

 gezicht

-0.43

 propagan

-0.43

POSITIVE LOGITS

:✨

1.16

<eos>

0.45

 ویکی‌آمباردا

0.43

endpush

0.41

 INTERESAR

0.40

 Autorisations

0.37

MemoryWarning

0.36

 Autorizaciones

0.36

比特派

0.36

uxxxx

0.35

Activations Density 0.000%

large sections of whitespace or empty content

identifiers and foreign words

No Comments

No Known Activations