INDEX

Explanations

technical terminology related to software and hardware compatibility

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

original electronic

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

issory

-0.34

zug

-0.33

 palha

-0.32

Krie

-0.32

lně

-0.31

InSection

-0.31

íně

-0.31

ambient

-0.31

subplot

-0.31

 laughing

-0.30

POSITIVE LOGITS

QMetaType

0.60

ThroughAttribute

0.59

 CreateTagHelper

0.50

compatibility

0.50

ніципалі

0.49

 Autorisations

0.49

🔕

0.48

 AssemblyTitle

0.48

 للاسماء

0.47

 AssemblyCulture

0.46

Activations Density 0.041%

technical terminology related to software and hardware compatibility

original electronic

No Comments

No Known Activations