INDEX

Explanations

terminology related to legal processes and examinations

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

cross-examination

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

<mask>

-0.54

laag

-0.50

 sondern

-0.44

señor

-0.43

hofer

-0.43

mist

-0.43

odyne

-0.42

 utafitiHapana

-0.42

 OnInit

-0.42

certain

-0.42

POSITIVE LOGITS

AxisAlignment

1.27

ChildScrollView

0.63

 cross

0.58

SharedDtor

0.57

complexContent

0.56

 crossed

0.55

 crossings

0.54

styleType

0.53

 Crossing

0.53

 crossing

0.52

Activations Density 0.001%

terminology related to legal processes and examinations

cross-examination

No Comments

No Known Activations