INDEX

Explanations

references to the court and legal terminology

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

after "court" or location words

np_acts-logits-general · gemini-2.0-flash

this court ruled

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 BrowserModule

-0.50

ViewFeatures

-0.45

wakili

-0.44

ThemeOverlay

-0.43

 grounded

-0.41

 engraçado

-0.41

VolleyError

-0.41

 ब्रेकडाउन

-0.41

 disconnected

-0.40

makeConstraints

-0.40

POSITIVE LOGITS

 للمعارف

0.46

 Вікіпе

0.43

DockStyle

0.43

HERE

0.43

 HERE

0.43

CodedInputStream

0.43

WindowConstants

0.41

 समीक्षाएं

0.40

WriteBarrier

0.40

 MonoBehaviour

0.40

Activations Density 0.658%

references to the court and legal terminology

after "court" or location words

this court ruled

No Comments

No Known Activations