INDEX

Explanations

technical and programming-related terms and structures

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

programming construct keywords

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

 nakalista

-0.57

\{\\

-0.56

 snippetHide

-0.51

 ModelExpression

-0.47

featureID

-0.44

complexContent

-0.44

 мәкал

-0.43

 الرياضيه

-0.43

addComponent

-0.41

 coincident

-0.41

POSITIVE LOGITS

sizePolicy

0.57

windowFixed

0.47



0.44

transQ

0.41

点此举报

0.41

InstrumentedTest

0.40

 HasFactory

0.39

softc

0.38

uParam

0.38

EndGlobalSection

0.37

Activations Density 0.106%

technical and programming-related terms and structures

programming construct keywords

No Comments

No Known Activations