Add optimality index and adaptive model wrapper

- OPTIMALITY_INDEX.md: catalog of all circuits tested with magnitude results
- prune.py: wrapper now tries multiple calling conventions (*args, list, tensor)

OPTIMALITY_INDEX.md

@@ -0,0 +1,69 @@
+# Optimality Index
+
+Results of exhaustive magnitude enumeration on threshold logic circuits.
+
+## Summary
+
+All circuits listed below have been verified via exhaustive enumeration. "Optimal" means no valid configuration exists at lower magnitude.
+
+## Single-Layer Gates (Linearly Separable)
+
+| Circuit | Inputs | Params | Optimal Mag | Solutions | Configs Tested |
+|---------|--------|--------|-------------|-----------|----------------|
+| threshold-not | 1 | 2 | 1 | 1 | 5 |
+| threshold-nor | 2 | 3 | 2 | 1 | 7 |
+| threshold-implies | 2 | 3 | 2 | 1 | 25 |
+| threshold-or | 2 | 3 | 3 | 1 | 25 |
+| threshold-nand | 2 | 3 | 3 | 1 | 25 |
+| threshold-and | 2 | 3 | 4 | 1 | 129 |
+| threshold-nor3 | 3 | 4 | 3 | 1 | 129 |
+| threshold-or3 | 3 | 4 | 4 | 1 | 321 |
+| threshold-nand3 | 3 | 4 | 5 | 1 | 681 |
+| threshold-and3 | 3 | 4 | 6 | 1 | 1,289 |
+| threshold-demux | 2 | 6 | 7 | 1 | 19,825 |
+| threshold-majority | 8 | 9 | 13 | 1 | 27,298,155 |
+
+## Multi-Layer Gates (Not Linearly Separable)
+
+| Circuit | Inputs | Params | Original Mag | Optimal Mag | Solutions | Reduction |
+|---------|--------|--------|--------------|-------------|-----------|-----------|
+| threshold-xor | 2 | 9 | 10 | 7 | 6 | 30% |
+| threshold-xnor | 2 | 9 | 9 | 7 | 2 | 22% |
+| threshold-mux | 3 | 11 | 10 | 7 | 4 | 30% |
+
+## Optimized Variants Created
+
+These repos contain the magnitude-optimal weights:
+
+- `threshold-xor-mag7` - 6 solutions at magnitude 7
+- `threshold-xnor-mag7` - 2 solutions at magnitude 7
+- `threshold-mux-mag7` - 4 solutions at magnitude 7
+
+## Pending / In Progress
+
+| Circuit | Params | Status |
+|---------|--------|--------|
+| threshold-halfadder | 12 | Running (expected optimal: 11) |
+| threshold-mod4 | 9 | Running |
+| threshold-biimplies | 9 | Not yet tested (same as XNOR) |
+| threshold-halfsubtractor | 12 | Not yet tested |
+
+## Methodology
+
+Exhaustive search enumerates all integer weight configurations by magnitude level (0, 1, 2, ...) until valid solutions are found. This guarantees the minimum magnitude is found.
+
+For circuits with >12 parameters, exhaustive search becomes impractical. Use evolutionary or simulated annealing instead.
+
+## Key Findings
+
+1. **Single-layer threshold gates have unique optimal representations.** All tested single-layer gates have exactly 1 solution at their optimal magnitude.
+
+2. **Multi-layer gates can have solution families.** XOR has 6 solutions at magnitude 7, organized into two structural families.
+
+3. **Non-linearly-separable functions benefit most from optimization.** XOR/XNOR/MUX achieved 22-30% magnitude reduction.
+
+4. **Optimal magnitude = sum of components for independent subnetworks.** Halfadder (XOR + AND) expected optimal is 7 + 4 = 11.
+
+## Last Updated
+
+2026-01-23

prune.py

@@ -581,7 +581,13 @@ class AdaptiveCircuit:
                         def wrapper(inputs, weights):
                             model = klass(weights)
                             args = [int(inputs[i]) for i in range(n)]
-                            result = model(*args)
+                            try:
+                                result = model(*args)
+                            except TypeError:
+                                try:
+                                    result = model(args)
+                                except TypeError:
+                                    result = model(torch.tensor(args, dtype=torch.float32))
                             return list(result) if isinstance(result, (list, tuple)) else [result]
                         return wrapper
                     return make_wrapper(attr, n_in)