INDEX

Explanations

occurrences of updates or announcements related to information or changes

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

update

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

TagMode

-0.41

saraba

-0.38

Anhalt

-0.38

 propOrder

-0.37

новништво

-0.36

étit

-0.36



-0.33

 Cous

-0.32

 уча

-0.32

 Royce

-0.32

POSITIVE LOGITS

 يتيمه

0.66

didSet

0.64

 nahilalakip

0.54

abestanden

0.54

帖最后由

0.53

updated

0.52

'{@

0.52

 updateUser

0.51

 updated

0.51

Aktualisiert

0.51

Activations Density 0.025%

occurrences of updates or announcements related to information or changes

update

No Comments

No Known Activations