INDEX

Explanations

phrases related to health conditions and patterns of substance use

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

chronic or repeated states

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

 ſei

-0.79

 للمعارف

-0.79

RTLR

-0.77

niſſe

-0.75

elemField

-0.74

bootstrapcdn

-0.74

 témoig

-0.73

 فريبيس

-0.71

AddTagHelper

-0.71

 imagui

-0.70

POSITIVE LOGITS

ๆ

0.38

 repeated

0.38

 chronic

0.36

繰り返

0.32

 Folge

0.31

 acumula

0.29

 costumbre

0.29

 persistent

0.29

Repeated

0.29

 kebiasaan

0.29

Activations Density 0.059%

phrases related to health conditions and patterns of substance use

chronic or repeated states

No Comments

No Known Activations