INDEX

Explanations

phrases related to legal, regulatory, or procedural contexts

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

preceding "of," "as," or "in"

np_acts-logits-general · gemini-2.0-flash

at least / from the viewpoint

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

 Usaha

-0.38

 instead

-0.37

 Constitucional

-0.37

 Dabei

-0.37

 statt

-0.35

 Statt

-0.35

 McCra

-0.35

 Maßnahmen

-0.34

 directamente

-0.33

 pájaros

-0.32

POSITIVE LOGITS

 CreateTagHelper

0.67

URLException

0.63

Filmografia

0.63

 فريبيس

0.62

 tartalomajánló

0.61

 Signalez

0.60

NameInMap

0.59

sizeCache

0.58

GraphicsUnit

0.56

 AssemblyTitle

0.56

Activations Density 0.460%

phrases related to legal, regulatory, or procedural contexts

preceding "of," "as," or "in"

at least / from the viewpoint

No Comments

No Known Activations