INDEX

Explanations

instances of statistical or configuration data related to components in a system

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

technical xml attribute names

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

 pleaſure

-0.68

ſelf

-0.59

 faſt

-0.56

 ſon

-0.55

 houſe

-0.54

 ſte

-0.54

 Houſe

-0.54

 Conſ

-0.53

 Reſ

-0.52

 Anſ

-0.51

POSITIVE LOGITS

GEBURTSDATUM

0.56

#+#

0.53

ModelAdmin

0.52

EndContext

0.52

ChildScrollView

0.51

PerformLayout

0.50

AccessorTable

0.49

 Normdatei

0.48

帖最后由

0.44

 propOrder

0.44

Activations Density 0.058%

instances of statistical or configuration data related to components in a system

technical xml attribute names

No Comments

No Known Activations