INDEX

Explanations

phrases indicating efforts and improvements in a team context

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

user input indicator

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

Partly

-0.32

Shame

-0.30

 způso

-0.30

 partly

-0.29

 Partly

-0.29

parti

-0.28

 télévision

-0.28

ucz

-0.28

 garrafa

-0.28

Element

-0.27

POSITIVE LOGITS

تقاوى

0.66

endcsname

0.63

 NSCoder

0.62

]")]

0.61

 '\\;'

0.59

 الحره

0.58

+#+

0.57

expandindo

0.57

providedIn

0.57

 cherchés

0.57

Activations Density 0.021%

phrases indicating efforts and improvements in a team context

user input indicator

No Comments

No Known Activations