INDEX

Explanations

verbs associated with improvement, prevention, and investigation in various contexts

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

actions that improve or produce results

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

ByVersion

-0.54

Autoritní

-0.52

الحياه

-0.50

 AssemblyCompany

-0.48

 PeEnEo

-0.48

 صوتيه

-0.47

 ویکی‌پدیا

-0.47

RegistryLite

-0.47

"];

-0.47

 للاسماء

-0.46

POSITIVE LOGITS

TagMode

0.47

 noemen

0.43

 exécut

0.42

 judicia

0.42

keyDown

0.39

 trekken

0.39

 assistência

0.39

 voelen

0.38

 belangrij

0.36

KeyType

0.36

Activations Density 0.048%

verbs associated with improvement, prevention, and investigation in various contexts

actions that improve or produce results

No Comments

No Known Activations