INDEX

Explanations

references to specific coding or computational elements, particularly related to files and modules

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Ży

-0.45

Determine

-0.44

 fahr

-0.44

 Whyte

-0.43

Darius

-0.42

kundige

-0.42

 Fournier

-0.42

 sayang

-0.42

 Xunit

-0.42

mathbb

-0.42

POSITIVE LOGITS

MOC

0.94

Moc

0.89

Mok

0.88

moc

0.84

mocha

0.83

Mox

0.81

Mok

0.81

 Mocha

0.79

MOC

0.78

Moc

0.75

Activations Density 0.006%

references to specific coding or computational elements, particularly related to files and modules

No Comments

No Known Activations