INDEX

Explanations

terms related to repetition and consistency

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

Autoritní

-0.56

 Offisielt

-0.55

 AppBar

-0.53

adaptiveStyles

-0.53

Personendaten

-0.53

 насељу

-0.53

 informée

-0.50

MemoryWarning

-0.50

 CanadaChoose

-0.49

PickerController

-0.48

POSITIVE LOGITS

 repeated

0.88

 frequent

0.78

repeated

0.77

 Repeated

0.77

Repeated

0.71

 recurring

0.68

 repeatedly

0.67

 repetition

0.67

 recurrent

0.66

 periodic

0.66

Activations Density 0.395%

terms related to repetition and consistency

No Comments

No Known Activations