INDEX

Explanations

references related to embedding content or objects in digital formats

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

API, SDK, RSS, embed, sites

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

featureID

-0.54

ParallelGroup

-0.53

PerformLayout

-0.50

StandardCharsets

-0.50

GraphicsUnit

-0.50

ConstraintMaker

-0.48

modelBuilder

-0.47

EndProject

-0.46

 للمعارف

-0.46

 ब्रेकडाउन

-0.44

POSITIVE LOGITS

 Embed

0.47

setVerticalGroup

0.44

 Италијани

0.40

 EconPapers

0.38

embed

0.38

 embed

0.37

Embed

0.37

 jadx

0.36

 anytime

0.35

ByEmail

0.35

Activations Density 0.005%

references related to embedding content or objects in digital formats

API, SDK, RSS, embed, sites

No Comments

No Known Activations