suyashnpande/narrative-context-m2

Narrative Context Module 2

Transformer encoder (4L × 8H) over a 5-scene sliding window. Consumes scene embeddings from Module 1.

Input

5-scene window [t-4 … t] (past + current only). Per-scene feature vector: 304-d (M1 embedding + metadata).

7 Cross-Scene Heads

#	Head	Type	Output
1	scene_valence_continuous	regression	-1.0 to 1.0
2	tension_level	regression	1 to 10
3	arousal_level	regression	1 to 10
4	emotional_shift_trigger	binary	True / False
5	narrative_arc_position	5-class	Setup, Rising, Climax, Falling, Resolution
6	foreshadowing_type	4-class	None, Foreshadow, Payoff, Echo
7	transition_type	5-class	attacca, fade, segue, silence, cut