INDEX

Explanations

technical terms and measurements related to engineering, processes, and methodologies

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

quantitative data suitable

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

VELAND

-0.49

 Савезне

-0.48

SharedCtor

-0.48

IndentedString

-0.47

MergeFrom

-0.46

 <=",

-0.45

SequentialGroup

-0.44

tonode

-0.43

 فريبيس

-0.39

 Ceinture

-0.38

POSITIVE LOGITS

GOTREF

0.55

 BoxFit

0.53

__":

0.50

esModule

0.48

<![

0.45

 guaranteed

0.42

MLLoader

0.42

guar

0.40

improved

0.40

 tartalomajánló

0.39

Activations Density 0.444%

technical terms and measurements related to engineering, processes, and methodologies

quantitative data suitable

No Comments

No Known Activations