INDEX

Explanations

terms related to chemical interactions and potential safety hazards

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

accidental contacts and exposure

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

المشاركات

-0.44

ultat

-0.35

олові

-0.33

########.

-0.33

ImageContext

-0.33

GenerationType

-0.31

richTextPanel

-0.30

 πάντα

-0.30

copyWith

-0.29

delwed

-0.29

POSITIVE LOGITS

 EXPOSURE

0.54

Exposure

0.54

MemoryWarning

0.53

 accidentally

0.52

 NSCoder

0.51

 MotionEvent

0.51

 AttributeSet

0.50

Cyfeiriadau

0.50

Bibitem

0.50

exposed

0.49

Activations Density 0.055%

terms related to chemical interactions and potential safety hazards

accidental contacts and exposure

No Comments

No Known Activations