INDEX

Explanations

instances of the word "ed" and its variations, indicating a focus on edited materials or publications

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

ending in "ed", "ing", or "ate"

np_acts-logits-general · gemini-2.0-flash

management and operations

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 24-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_24/width_16k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.24.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

TypedDataSet

-0.70

الحياه

-0.68

 للاسماء

-0.68

 Wiktionnaire

-0.65

IsContent

-0.64

 '\\;'

-0.58

Tikang

-0.57

 فريبيس

-0.56

цездатний

-0.55

 الرياضيه

-0.53

POSITIVE LOGITS

 guided

0.45

led

0.44

 diriger

0.43

 oversaw

0.42

 overseen

0.41

 Leitung

0.41

 affairs

0.40

CloseOperation

0.40

 assisted

0.39

enação

0.39

Activations Density 0.450%

instances of the word "ed" and its variations, indicating a focus on edited materials or publications

ending in "ed", "ing", or "ate"

management and operations

No Comments

No Known Activations

instances of the word "ed" and its variations, indicating a focus on edited materials or publications

ending in "ed", "ing", or "ate"

management and operations

No Comments

No Known Activations