INDEX

Explanations

references to structured data or processes, particularly those involving technical or programming contexts

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

 Administrativna

-0.69

 ExecuteAsync

-0.69

 ویکی‌پدی

-0.68

 HttpNotFound

-0.66

 autorytatywna

-0.64

 informée

-0.64

ScopeManager

-0.62

AsUp

-0.60

 للاسماء

-0.56

 TextAppearance

-0.53

POSITIVE LOGITS

 located

0.46

 nationales

0.42

 wherever

0.42

 Located

0.42

Located

0.41

 coastal

0.40

Locations

0.40

 Northern

0.36

DockStyle

0.35

southern

0.34

Activations Density 0.105%

references to structured data or processes, particularly those involving technical or programming contexts

No Comments

No Known Activations