INDEX

Explanations

identifiers and structured data elements in code or technical documents

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

mathematical expressions and references

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

 صوتيه

-0.72

transQ

-0.71

 resourceCulture

-0.67

 Chwiliwch

-0.67

 Italijanski

-0.66

 فريبيس

-0.61

 étoient

-0.58

verifyException

-0.57

InputBorder

-0.57

 HasFactory

-0.57

POSITIVE LOGITS

heitsbild

0.37

NamedQueries

0.32

sstream

0.32

 ок

0.30

vd

0.28

çon

0.28

Dy

0.28

 used

0.28

adox

0.28

 yêu

0.27

Activations Density 0.003%

identifiers and structured data elements in code or technical documents

mathematical expressions and references

No Comments

No Known Activations