INDEX

Explanations

updates and enhancements regarding rules and policies

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 17-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_17/width_16k/average_l0_73

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.17.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

ьаж

-0.69

 Wikimedijinoj

-0.62

 utafitiHapana

-0.60

GEBURTSDATUM

-0.55

UnusedPrivate

-0.54

TagMode

-0.53

Tembelea

-0.52

 NSCoder

-0.50

 nonUne

-0.49

 pinulongan

-0.49

POSITIVE LOGITS

 improved

0.55

improved

0.51

Easier

0.49

 easier

0.47

Improved

0.47

easier

0.47

amélioration

0.46

 verbess

0.45

 smoother

0.45

 increased

0.44

Activations Density 0.445%

updates and enhancements regarding rules and policies

No Comments

No Known Activations