INDEX

Explanations

references to awards and winners in a competition context

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

awards and selections

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_24/width_131k/average_l0_97

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.24.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

 '\\;'

-0.52

 tartalomajánló

-0.51

بوابة

-0.47

ArrowToggle

-0.45

 MainAxisSize

-0.44

Tembelea

-0.41

AnchorStyles

-0.41

FontOfSize

-0.40

/**

-0.40

Спољашње

-0.39

POSITIVE LOGITS

:✨

0.54

 Komunikasi

0.50

참고

0.50

selected

0.48

 selected

0.47

 wybran

0.44

 Keuangan

0.43

GEBURTSDATUM

0.43

 Kecil

0.43

 dipilih

0.43

Activations Density 0.009%

references to awards and winners in a competition context

awards and selections

No Comments

No Known Activations