id
stringlengths
14
14
type
stringclasses
21 values
name_nodes
stringlengths
2
36
mean_nodes
stringlengths
2
11.7k
frame_nodes
stringlengths
2
216
raw_content
stringlengths
1
560
0x27b88ea58118
base_name
[]
[]
[]
2. Two Paths to Expand Context
0x952c62e843b3
principle
["0x27b88ea58118"]
[]
["0x501d59a1e46c", "0x15fbb59966b0", "0x90563024aeb2"]
2. Two Paths to Expand Context When context is insufficient, there are exactly two ways to expand it: **1. Imagination (endogenous)** > Use what is already known to infer what is not yet known. > Generate possible cases without requiring additional input from the world. **2. Exploration (exogenous)** > Acquire additi...
0x9256b4ecbe34
base_name
[]
[]
[]
3. The Convergence Loop
0xda008ef7483f
principle
["0x9256b4ecbe34"]
[]
["0x501d59a1e46c", "0x15fbb59966b0", "0xfbaec867b0e9", "0x90563024aeb2"]
3. The Convergence Loop The two paths are not independent — they loop: ``` context insufficient → [1] imagine: generate cases from what is known → [2] explore: observe to confirm or refute → gap narrows → context updated → sufficient → collapse → insufficient → return to [1] with updated context ``` Converg...
0xe8f469964683
base_name
[]
[]
[]
4. Meaning Depends on Frame
0xdde0a531ff5f
principle
["0xe8f469964683"]
[]
["0x501d59a1e46c", "0x15fbb59966b0", "0x0ad428c0a634"]
4. Meaning Depends on Frame ``` M = f(N, F) ``` - **N**: name / identifier of a concept - **F**: the set of currently active context — the frame - **M**: meaning — only valid within that specific F Same N, different F → different M. There is no absolute meaning. Meaning is a relation between concept and context. Col...
0x5acb21c154f4
base_name
[]
[]
[]
5. Memory is Layered by Rate of Change
0x2da5380afcb2
principle
["0x5acb21c154f4"]
[]
["0x501d59a1e46c", "0x91515f441344", "0xfbe37abdbc60", "0x700ff6e270fd"]
5. Memory is Layered by Rate of Change Not all context changes at the same speed. Slower layers are not reset by faster layers. ``` invariant → cross-domain, never changes quasi-invariant → learned gradually, stable after many observations spatial / episodic → changes per situation momentary → exi...
0x46c9a1c3a457
base_name
[]
[]
[]
6. Gap is the Driver of Learning
0xb8f06727c897
principle
["0x46c9a1c3a457"]
[]
["0x501d59a1e46c", "0x15fbb59966b0", "0x90563024aeb2"]
6. Gap is the Driver of Learning Gap = the distance between imagination and observation. Gap is not an error — gap is a signal. No gap → no learning. Large persistent gap → fastest learning. ``` gap appears → classify: where in the loop does the gap live? → generate hypothesis to fill the gap → test via explora...
0x0f6c4645b0ce
base_name
[]
[]
[]
7. Multi-channel Input — Collect First, Filter Later
0x4408ea16d4e2
principle
["0x0f6c4645b0ce"]
[]
["0x501d59a1e46c", "0x852eecb66da1", "0xfbe37abdbc60", "0x0ad428c0a634", "0xfbaec867b0e9", "0x16f9f618ef6e"]
7. Multi-channel Input — Collect First, Filter Later The human brain renders a world model from at least 5 parallel channels. No channel is "primary" — each provides a type of context that no other channel can replace. **Principle:** > You cannot know in advance which channel is meaningful in this situation. > Collec...
0x8567f1c3b538
base_name
[]
[]
[]
8. Three Object Classes in Any Game
0x74b2e67eea2b
principle
["0x8567f1c3b538"]
[]
["0x501d59a1e46c", "0x91515f441344", "0xfbe37abdbc60", "0x0ad428c0a634", "0x700ff6e270fd"]
8. Three Object Classes in Any Game Every game world has exactly 3 types of entities: **1. Avatar Player (self)** > The entity you control. Moves in response to your actions. > You do NOT "become" the player — you OBSERVE and CONTROL it from outside (3rd person perspective). > This means: you can see the player's coor...
0xa8bcddb17909
base_name
[]
[]
[]
9. Context Query Protocol — Questions Before Actions
0xf09b98a55109
principle
["0xa8bcddb17909"]
[]
["0x501d59a1e46c", "0xfbe37abdbc60", "0x15fbb59966b0", "0xfbaec867b0e9", "0x16f9f618ef6e", "0x90563024aeb2"]
9. Context Query Protocol — Questions Before Actions > The quality of a decision is bounded by the quality of the context that produced it. > Context does not arrive automatically — it must be QUERIED. > Each query = attention directed at a specific sensor or memory channel. > No query = no context from that channel = ...
0xa434b76bf070
base_name
[]
[]
[]
10. Universal Game Solving — Interaction Chain
0x16bc590112dd
principle
["0xa434b76bf070"]
[]
["0x501d59a1e46c", "0x91515f441344", "0xfbe37abdbc60", "0x15fbb59966b0", "0x0ad428c0a634", "0x700ff6e270fd", "0x90563024aeb2"]
10. Universal Game Solving — Interaction Chain > This principle applies to ALL games, ALL levels, ALL genres. > It is the invariant structure of how difficulty scales. **The simplest game:** ``` A → B Player → Goal ``` **How difficulty increases — the ONLY way:** ``` A → [interact O1] → B A → [interact O1] → [interac...
0x1fe7f13596e2
base_name
[]
[]
[]
11. Think Before Look — Superposition of Possibilities
0xaad993208b13
principle
["0x1fe7f13596e2"]
[]
["0x501d59a1e46c", "0x91515f441344", "0xfbe37abdbc60", "0x700ff6e270fd", "0xfbaec867b0e9", "0x16f9f618ef6e", "0x90563024aeb2"]
11. Think Before Look — Superposition of Possibilities > Everything exists in superposition until input collapses it. > Cognition is NOT: see → think → act. > Cognition IS: think (imagine possibilities) → see (input collapses) → act (on collapsed reality). **Before receiving ANY new input, the mind must IMAGINE what i...
0xb220c9010b97
base_name
[]
[]
[]
12. Color Gradient = Map Topology
0x8c5db08bec89
principle
["0xb220c9010b97"]
[]
["0x501d59a1e46c", "0x852eecb66da1", "0xfbe37abdbc60", "0x15fbb59966b0"]
12. Color Gradient = Map Topology > Before detecting objects, read the MAP itself from color gradients. > Bright pixels = paths. Dark pixels = walls. The brightness pattern IS the road network. In ARC-3 games: ``` color 3 (dark gray, bright relative to walls) = corridor = CAN walk color 4 (near black) ...
0x6440644d0596
base_name
[]
[]
[]
13. Bipolar Field Map — Object as Anchor Between Invariant and Variant
0x9fe76cd76d71
principle
["0x6440644d0596"]
[]
["0x501d59a1e46c", "0x91515f441344", "0x15fbb59966b0", "0x700ff6e270fd", "0x16f9f618ef6e", "0x90563024aeb2"]
13. Bipolar Field Map — Object as Anchor Between Invariant and Variant > Every object sits at the CENTER between two poles. > Properties of the object are VALUE NODES that orbit between the object and their pole. > This is the fundamental data structure for representing knowledge. **Structure:** ``` [INVARIANT POLE] ←...
0x2c95a4388a18
base_name
[]
[]
[]
14. The Fundamental Triad — Player, Door, Object
0x899d8f1d9217
principle
["0x2c95a4388a18"]
[]
["0x501d59a1e46c", "0xfbe37abdbc60", "0x700ff6e270fd"]
14. The Fundamental Triad — Player, Door, Object > Every game is exactly 3 things: > 1. PLAYER (constant) — always exists, always controllable, always wants to reach door. > 2. DOOR (constant) — always exists, always the goal, always requires conditions to enter. > 3. OBJECTS (variable) — everything between player and ...
0xf02ade742b38
base_name
[]
[]
[]
15. Node Flattening — List All, Then Fold by Shared Values
0x583103bd0498
principle
["0xf02ade742b38"]
[]
["0x501d59a1e46c", "0xfbe37abdbc60", "0x15fbb59966b0"]
15. Node Flattening — List All, Then Fold by Shared Values > When there are many objects between Player and Door, the chain gets complex. > Solution: flatten everything into individual value nodes, then FOLD by shared values. **Step 1: Flatten** Every object → list ALL its properties as individual nodes on the Player→...
0x4094b9ec7038
base_name
[]
[]
[]
16. Test Unknown Objects Using ALL Known Mechanisms
0xe5a7501fd97f
principle
["0x4094b9ec7038"]
[]
["0x501d59a1e46c", "0xfbe37abdbc60"]
16. Test Unknown Objects Using ALL Known Mechanisms > When you discover a new object, do NOT test it only one way. > Test it using EVERY mechanism you have already learned from other objects. **The mistake:** - Rotator triggers by "pass through" (not "stand next to") - Yellow square tested by "stand next to + press di...
0xa8d326b5e850
base_name
[]
[]
[]
17. Fractal Decomposition of Input — Invariant Extraction at Every Scale
0x6546dcaa6bdc
principle
["0xa8d326b5e850"]
[]
["0x501d59a1e46c", "0x91515f441344", "0xfbe37abdbc60", "0x700ff6e270fd", "0x16f9f618ef6e"]
17. Fractal Decomposition of Input — Invariant Extraction at Every Scale > Sensor collects raw signal. I.P PROCESSES it by extracting invariant at every scale. > Never test specific cases. Extract the RULE that covers all cases. **Method: recursive split of VARIANT into (invariant_small + variant_small)** ``` Layer 1...
0x9c3bc2783720
base_name
[]
[]
[]
18. Binary Classification First — The Universal Split
0x92d4031822ae
principle
["0x9c3bc2783720"]
[]
["0x501d59a1e46c", "0x91515f441344", "0xfbe37abdbc60", "0x16f9f618ef6e"]
18. Binary Classification First — The Universal Split > Before any game-specific logic, split ALL pixels into exactly 2 groups. > Then split again. 2 → 3 → done. Universal. Works on any 2D game. **Step 1: Binary split** ``` [ALL PIXELS] / \ [RESPONSIVE] [NON-RESPONSIVE] (delta > 0) (d...
0x85636c9b5780
base_name
[]
[]
[]
19. Reverse Trace — Learn Mechanism from Solution, Not from Trial
0x4e5434589a9a
principle
["0x85636c9b5780"]
[]
["0x501d59a1e46c", "0xfbe37abdbc60", "0x15fbb59966b0", "0x16f9f618ef6e", "0x90563024aeb2"]
19. Reverse Trace — Learn Mechanism from Solution, Not from Trial > If you have both the start (A) and the end (Z), don't learn forward A→Z. > Learn BACKWARD Z→A. The backward trace reveals the mechanism. **Forward learning** (A→Z): try action → observe → try another → slow, wasteful, may never converge. **Backward t...
0xb2abb3f544f2
base_name
[]
[]
[]
20. Empirical Mechanism Distribution — What ARC Actually Requires
0x96def52c02a4
principle
["0xb2abb3f544f2"]
[]
["0x501d59a1e46c", "0x15fbb59966b0", "0x0ad428c0a634", "0xfbaec867b0e9", "0x16f9f618ef6e"]
20. Empirical Mechanism Distribution — What ARC Actually Requires > From reverse-tracing 400 ARC classic tasks (2874 observations): > The data tells you what matters. Not theory — data. **Binary tree of ALL mechanisms found:** ``` L0: SHAPE CHANGES(487) vs SAME SHAPE(815) │ ├── SHAPE CHANGES (37%) │ ├── expand(121)...
0x0cf670172556
base_name
[]
[]
[]
21. Three Chain Orders — The Only Structures That Exist
0xf0f86879690d
principle
["0x0cf670172556"]
[]
["0x501d59a1e46c", "0x91515f441344", "0xfbe37abdbc60", "0x700ff6e270fd"]
21. Three Chain Orders — The Only Structures That Exist > From 400 ARC tasks reverse-traced: only 3 chain orderings appear. Not 6. Not arbitrary. > Chain STRUCTURE is invariant. Chain CONTENT is variant (fractal sub-chains inside). **The 3 orders (empirical, from 1302 observations):** ``` Order 1 (50%): WHY → WHERE →...
0xd0cf6574e5d2
base_name
[]
[]
[]
22. Five Grounding Primitives — Binary Is Not Enough
0xa56e17c296c2
principle
["0xd0cf6574e5d2"]
[]
["0x501d59a1e46c", "0x91515f441344", "0x852eecb66da1", "0xfbe37abdbc60", "0x15fbb59966b0", "0xed0b40836a29"]
22. Five Grounding Primitives — Binary Is Not Enough > Binary (P18) covers most classification but cannot represent everything alone. > Minimum grounding set = 5 irreducible primitives: | # | Primitive | Mechanism | Covers | |---|-----------|-----------|--------| | G1 | **Binary** | A ↔ B | distinction, existence, K/U...
0x23f98228b589
base_name
[]
[]
[]
23. Epistemic vs Exploit Action
0xe9aeabe8eb71
principle
["0x23f98228b589"]
[]
["0x501d59a1e46c", "0x852eecb66da1"]
23. Epistemic vs Exploit Action > When rules unknown: best action = maximize INFORMATION GAIN, not reward. ``` Epistemic: "what does this action REVEAL about the system?" Exploit: "what does this action GET me toward goal?" ``` Switch from epistemic to exploit when confidence > threshold. Not by step count — by con...
0x69c49da6d806
base_name
[]
[]
[]
24. NMF — Name, Meaning, Frame
0xab6899eb90f6
principle
["0x69c49da6d806"]
[]
["0x501d59a1e46c", "0x15fbb59966b0", "0x0ad428c0a634", "0x700ff6e270fd"]
24. NMF — Name, Meaning, Frame > Every piece of information is filtered through 3 simultaneous lenses: ``` Name = f(Meaning, Frame) ``` - **Name** — label in current context (can change per frame) - **Meaning** — stable concept across contexts (transfers) - **Frame** — reference system used to interpret (can switch) ...
0xe7fe09267d7b
base_name
[]
[]
[]
25. E = Superpose(K, U) — Existence as Grounded Known + Structured Unknown
0x06a09513b910
principle
["0xe7fe09267d7b"]
[]
["0x501d59a1e46c", "0x91515f441344", "0x700ff6e270fd", "0x0f437799202f"]
25. E = Superpose(K, U) — Existence as Grounded Known + Structured Unknown > Classical: A = constant + variable. > General: E = Superpose(K, U). - **K** = grounded known anchor (stabilized, not just "constant") - **U** = structured unknown (NOT noise — unresolved structured presence) ``` U = V(K, Δ₁, Δ₂, ..., Δn) Δ...
0x310768bb9aef
base_name
[]
[]
[]
26. Grounded Node = [G, T, S, W, X]
0x9e9f14b0d8ff
principle
["0x310768bb9aef"]
[]
["0x501d59a1e46c", "0x15fbb59966b0"]
26. Grounded Node = [G, T, S, W, X] > A node is not a point in a graph. It is a grounded entity in an existence field. ``` N = [G, T, S, W, X] G = Grounding — primordial axes coordinates, ontological anchor T = Time trace — when created, when last activated S = State — current expression (DERIVED, no...
0x59e09db80c55
base_name
[]
[]
[]
27. Five-Layer Architecture Stack
0x3ff163dd704f
principle
["0x59e09db80c55"]
[]
["0x501d59a1e46c", "0x91515f441344", "0x15fbb59966b0", "0x0ad428c0a634", "0x0f437799202f"]
27. Five-Layer Architecture Stack > From substrate to compression — the full grounded system: ``` Layer 0: Existence Core (E) — substrate containing everything Layer 1: Primordial Axes (P) — binary grounding poles (K↔U, etc.) Layer 2: Grounded Nodes (N) — entities anchored to at least 1 axis Lay...
0x9905d1439e93
base_name
[]
[]
[]
28. Frame Adjudicator — ACCEPT / REJECT / PARALLEL
0x965f29b86147
principle
["0x9905d1439e93"]
[]
["0x501d59a1e46c", "0xfbe37abdbc60", "0x16f9f618ef6e"]
28. Frame Adjudicator — ACCEPT / REJECT / PARALLEL > Before processing any input, evaluate the frame: ``` ACCEPT — frame consistent with observation → proceed REJECT — frame contradicts observation → reframe from internal model PARALLEL — insufficient evidence → maintain multiple frames → choose action ...
0x3faa3c845140
base_name
[]
[]
[]
29. World Model = 5 Primitives
0xb09efa80ec61
principle
["0x3faa3c845140"]
[]
["0x501d59a1e46c", "0x852eecb66da1"]
29. World Model = 5 Primitives > To represent ANY interactive world: 1. **Spatial occupancy** — where objects are, what space they fill 2. **Movement vector** — how objects move 3. **Static vs dynamic** — background vs actors 4. **Scalar state** — values that change (health, energy, count) 5. **Contact effects** — wha...
0xf4afd2b71e52
base_name
[]
[]
[]
30. Scaffold Architecture — IPOD + VEG
0x90ed35855707
principle
["0xf4afd2b71e52"]
[]
["0x501d59a1e46c", "0x91515f441344", "0xfbe37abdbc60", "0x700ff6e270fd", "0xfbaec867b0e9", "0x16f9f618ef6e", "0x90563024aeb2"]
30. Scaffold Architecture — IPOD + VEG The principles above are implemented through two orthogonal structures: **IPOD = the body (invariant structure)** ``` I (Input) — sensors that collect all channels (P7) P (Process) — brain that runs convergence loop (P3) using query protocol (P9) O (Output) — actuators that ex...
0xb4c612ede35e
base_name
[]
[]
[]
31. Collapse Condition = Context Count, Not Numeric Threshold
0x4e4f4731f378
principle
["0xb4c612ede35e"]
[]
["0x501d59a1e46c", "0x91515f441344", "0x852eecb66da1", "0xfbe37abdbc60", "0x15fbb59966b0", "0x0ad428c0a634", "0x700ff6e270fd", "0xfbaec867b0e9", "0x16f9f618ef6e", "0x90563024aeb2"]
31. Collapse Condition = Context Count, Not Numeric Threshold > A chain does not collapse by exceeding a fixed numeric threshold. > A chain collapses when the NUMBER OF MATCHING CONTEXT CLUES is maximal. **Wrong approach:** ``` score = Σ (power_i × ratio_i) ← fixed numbers, same for every input if score > thresho...
0xc7484593ee95
base_name
[]
[]
[]
32. Frame Switch Eliminates U
0x969914de2ae2
principle
["0xc7484593ee95"]
[]
["0x501d59a1e46c", "0x91515f441344", "0xfbe37abdbc60", "0xfbaec867b0e9", "0x16f9f618ef6e"]
32. Frame Switch Eliminates U > When stuck, the problem is not lack of compute — it is lack of frame. > Switch frame → see new K/U split → new information → U shrinks → closer to collapse. Every observation is filtered through a frame. Same data, different frame → different K and different U. When all actions within t...
0x54ea06e69006
base_name
[]
[]
[]
33. Reduce to 2^10 Before Brute Force
0x7d079b652339
principle
["0x54ea06e69006"]
[]
["0x501d59a1e46c", "0x91515f441344", "0xfbe37abdbc60", "0x15fbb59966b0"]
33. Reduce to 2^10 Before Brute Force > Never brute force more than ~1000 combinations. If search space > 2^10, STOP and reduce first. > Human cannot enumerate 4^31. Neither should agent. Add input to shrink U. **Reduction chain (each step cuts exponentially):** ``` RAW: N_objects ^ N_clicks (e.g., 4...
0x3bb88ce3537b
base_name
[]
[]
[]
34. Metaphor = Structure Transfer Across Domains
0xa1a06e5855a4
principle
["0x3bb88ce3537b"]
[]
["0x501d59a1e46c", "0x15fbb59966b0", "0x0ad428c0a634"]
34. Metaphor = Structure Transfer Across Domains > When stuck in domain A, find domain B with ISOMORPHIC structure. Solution in B transfers to A. > This is P24 (NMF) at macro scale: same Meaning, different Name, different Frame. Metaphor is not decoration — it is the mechanism of cross-domain intelligence. ``` Domain...
0x844970691d53
base_name
[]
[]
[]
35. Gap + Metaphor = Universal Reasoning Tool
0xf37c1da19133
principle
["0x844970691d53"]
[]
["0x501d59a1e46c", "0x91515f441344", "0xfbe37abdbc60", "0x15fbb59966b0", "0x0ad428c0a634", "0x700ff6e270fd", "0xfbaec867b0e9", "0x16f9f618ef6e", "0x90563024aeb2"]
35. Gap + Metaphor = Universal Reasoning Tool > Gap is not just P6 (learning signal). Gap is the UNIVERSAL MEASURE of difference when comparing any two things. > Metaphor is not just P34 (structure transfer). Metaphor provides GROUNDING for the unknown. > Combined: metaphor gives you a complete structure to compare aga...
0x7960621cdb5b
base_name
[]
[]
[]
01_TOB_Framework
0x147963c74a2b
category
["0x7960621cdb5b"]
[]
["0x501d59a1e46c"]
01_TOB_Framework
0x107412ad6cd0
base_name
[]
[]
[]
04_Others
0xa68ce18afd9b
category
["0x107412ad6cd0"]
[]
["0x501d59a1e46c"]
04_Others
0x737f4396e5ce
base_name
[]
[]
[]
05_System_Prompt
0x64d9f5ce0a51
category
["0x737f4396e5ce"]
[]
["0x501d59a1e46c"]
05_System_Prompt
0x1f8a84a372f9
base_name
[]
[]
[]
Danh_Nghia_He
0x59442f906471
keyword
["0x8c7c32544486"]
[]
[]
flow
0x8c7c32544486
base_name
[]
[]
[]
flow
0xd7174fe4b292
keyword
["0xa21faa3941b0"]
[]
[]
output
0xa21faa3941b0
base_name
[]
[]
[]
output
0x8e3d76e6df6f
keyword
["0x4e2484008abe"]
[]
[]
graph
0x4e2484008abe
base_name
[]
[]
[]
graph
0x83b5a2c2f920
keyword
["0xa0abd9df4e8e"]
[]
[]
learning
0xa0abd9df4e8e
base_name
[]
[]
[]
learning
0xc6f62c6de314
keyword
["0x12faded5c82a"]
[]
[]
invariant
0x12faded5c82a
base_name
[]
[]
[]
invariant
0x7151fc571443
keyword
["0x4416ad448cb6"]
[]
[]
node
0x4416ad448cb6
base_name
[]
[]
[]
node
0xcc88b6fef577
keyword
["0x8cab01d21b72"]
[]
[]
input
0x8cab01d21b72
base_name
[]
[]
[]
input
0xd108758a5fb5
keyword
["0xee9361d7806d"]
[]
[]
variant
0xee9361d7806d
base_name
[]
[]
[]
variant
0x92eab4e046c9
keyword
["0x533a14a6b55f"]
[]
[]
edge
0x533a14a6b55f
base_name
[]
[]
[]
edge
0x7bf3334e8663
keyword
["0xf6f268ccddf6"]
[]
[]
threshold
0xf6f268ccddf6
base_name
[]
[]
[]
threshold
0x1f64900026eb
keyword
["0xf5df57509a5c"]
[]
[]
data
0xf5df57509a5c
base_name
[]
[]
[]
data
0x2fe776bc92e5
keyword
["0x34a31d7ffc86"]
[]
[]
ipod
0x53070d9c0f82
keyword
["0xebcd41d11f3d"]
[]
[]
sao ch
0xebcd41d11f3d
base_name
[]
[]
[]
sao ch