INDEX

Explanations

instances of people, particularly their names and actions related to their careers or roles

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

name followed by a verb

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

 kasarigan

-0.55

:✨

-0.54

TypedDataSet

-0.49

 kaarangay

-0.49

 propOrder

-0.46

 الرياضيه

-0.45

oa̍t

-0.44

ineſs

-0.43

цездатний

-0.42

Thể

-0.41

POSITIVE LOGITS

 disambiguazione

0.52



0.51

CloseOperation

0.48

 Económica

0.48

</thead>

0.42

BindingResult

0.42

niająca

0.42

 Flask

0.42

__).

0.41

 człowie

0.41

Activations Density 0.014%

instances of people, particularly their names and actions related to their careers or roles

name followed by a verb

No Comments

No Known Activations