Index/games
House · game designers, interactive artists, XR creators
🎮

Game Design & Interactive Media

100
Entries
4
Primitives

Showing 100 of 100 entries

action game readabilitya-0

Boss Telegraph LoRA

Fine-tune Qwen/Qwen3-8B on 90 boss move telegraphs so it writes readable attack windups for combat designers.

SFT · Instruction
narrative-system cohesionh-1

Ludonarrative Patch

Fine-tune Qwen/Qwen3.5-4B on 70 plot-mechanic fixes so writers spot system-story conflicts early.

SFT · Instruction
generative level grammars-2

Procedural Seed Words

Fine-tune Qwen/Qwen3-8B on 80 seed-to-rule maps so it suggests PCG parameters from a theme.

SFT · Instruction
input tuning critiques-3

Controller Feel Notes

Fine-tune Qwen/Qwen3.5-4B on 60 feel critiques so it rates gamepad mapping snappiness.

SFT · Instruction
stylized rendering briefsa-4

Shader Poetry LoRA

Fine-tune Qwen/Qwen3-8B on 75 art-direction briefs so it writes shader intent for technical artists.

SFT · Instruction
tutorial pacingh-5

Onboarding Ramp Smith

Fine-tune Qwen/Qwen3.5-4B on 65 tutorial beats so it drafts gentle first-minute learning curves.

SFT · Instruction
immersive interfacer-6

Diegetic UI Whisper

Fine-tune Qwen/Qwen3-8B on 70 UI-in-world fixes so it proposes diegetic HUD replacements.

SFT · Instruction
virtual economy tuningr-7

Economy Balance Pair

Fine-tune Qwen/Qwen3.5-9B on 85 currency logs so it predicts inflation from drop rates.

SFT · Instruction
comfort-aware movementb-8

XR Locomotion Lab

Fine-tune Qwen/Qwen3.5-4B on 60 VR move tests so it recommends nausea-safe locomotion.

SFT · Instruction
encounter compositionf-9

Enemy Variety Chef

Fine-tune Qwen/Qwen3-8B on 90 foe mixes so it assembles balanced enemy squads by biome.

SFT · Instruction
audio direction text-10

Sound Cue Script

Fine-tune Qwen/Qwen3.5-4B on 75 cue notes so it writes SFX briefs from gameplay moments.

SFT · Instruction
branching structure-11

Quest Tree Pruner

Fine-tune Qwen/Qwen3.5-9B on 80 quest graphs so it trims bloated narrative branches.

SFT · Instruction
palette emotion-12

Color Mood Lens

Fine-tune Qwen/Qwen3-8B on 70 palette moods so it maps scenes to emotional color scripts.

SFT · Instruction
juice and feedback-13

Game Feel Microtips

Fine-tune Qwen/Qwen3.5-4B on 65 juice notes so it suggests screen-shake and hit-stop values.

SFT · Instruction
extensibility docs-14

Mod Hook Writer

Fine-tune Qwen/Qwen3-8B on 85 hook docs so it authors modding API examples from specs.

SFT · Instruction
session rhythm-15

Pacing Curve Smith

Fine-tune Qwen/Qwen3.5-9B on 80 session curves so it plots tension arcs for levels.

SFT · Instruction
inclusive design-16

Accessibility Pass

Fine-tune Qwen/Qwen3.5-4B on 70 a11y fixes so it flags colorblind and motor barriers.

SFT · Instruction
checkpoint narrative-17

Save State Poet

Fine-tune Qwen/Qwen3-8B on 60 checkpoint lines so it writes resonant save-screen text.

SFT · Instruction
fighting game metrics-18

Frame Data Scribe

Fine-tune Qwen/Qwen3.5-9B on 90 frame tables so it explains advantage frames in plain words.

SFT · Instruction
environment storytelling-19

Ambient World Hum

Fine-tune Qwen/Qwen3.5-4B on 75 world notes so it drafts ambient lore from room geometry.

SFT · Instruction
progression wording-20

Rewards Text Forge

Fine-tune Qwen/Qwen3-8B on 85 reward lines so it writes motivating unlock descriptions.

SFT · Instruction
retro asset grammar-21

Pixel Grid Rune

Fine-tune Qwen/Qwen3.5-4B on 70 tile rules so it describes cohesive pixel tile sets.

SFT · Instruction
playtest insight-22

Beta Signal Lens

Fine-tune Qwen/Qwen3.5-9B on 80 playtest notes so it summarizes friction from session logs.

SFT · Instruction
deckbuilder balance-23

Card Synergy Bard

Fine-tune Qwen/Qwen3-8B on 90 card pairs so it names synergy and counter relationships.

SFT · Instruction
reveal editing-24

Trailer Beat Maker

Fine-tune Qwen/Qwen3.5-4B on 65 trailer cuts so it drafts beat sheets from game footage tags.

SFT · Instruction
procedural level topologye-0

Level Graph Scribe

Fine-tune Qwen/Qwen3-8B on 80 (brief, graph DSL) pairs so it emits valid node-link level maps from plain text.

SFT · Structured Output
branching narrativer-1

Dialogue Tree Forger

Fine-tune Qwen/Qwen3.5-4B on 70 (synopsis, YAML tree) pairs so it outputs valid dialogue branches from a story seed.

SFT · Structured Output
color grading rulesh-2

Palette Rule Smith

Fine-tune Qwen/Qwen3.5-4B on 60 (mood, hex JSON) pairs so it returns constrained palettes from a feeling brief.

SFT · Structured Output
mission state machinesr-3

Quest Logic Weaver

Fine-tune Qwen/Qwen3-8B on 90 (prompt, quest JSON) pairs so it emits valid state machines from a quest idea.

SFT · Structured Output
material shader configr-4

Shader Param Carver

Fine-tune Qwen/Qwen3.5-4B on 50 (look, uniform JSON) pairs so it outputs shader uniform sets from a style note.

SFT · Structured Output
rhythm game chartsr-5

Beat Drop Mapper

Fine-tune Qwen/Qwen3-8B on 65 (song, chart JSON) pairs so it returns timed note charts from a track description.

SFT · Structured Output
behavior schemasr-6

NPC Rule Drafter

Fine-tune Qwen/Qwen3.5-9B on 75 (role, behavior JSON) pairs so it emits AI behavior specs from a character brief.

SFT · Structured Output
onboarding flowsm-7

Tutorial Step Loom

Fine-tune Qwen/Qwen3.5-4B on 55 (game, steps JSON) pairs so it outputs tutorial step lists from a mechanic summary.

SFT · Structured Output
effect triggersr-8

VFX Event Coder

Fine-tune Qwen/Qwen3-8B on 70 (action, VFX JSON) pairs so it returns effect event configs from a combat note.

SFT · Structured Output
world codex structr-9

Lore Index Binder

Fine-tune Qwen/Qwen3.5-4B on 80 (lore dump, codex JSON) pairs so it outputs linked lore entries from raw notes.

SFT · Structured Output
control bindings-10

Input Map Etcher

Fine-tune Qwen/Qwen3.5-4B on 45 (device, bindings JSON) pairs so it emits control maps from a platform brief.

SFT · Structured Output
persistence format-11

Save Schema Mason

Fine-tune Qwen/Qwen3-8B on 60 (state, save JSON) pairs so it outputs save-file schemas from a game spec.

SFT · Structured Output
resource curves-12

Economy Balance Slate

Fine-tune Qwen/Qwen3.5-9B on 85 (loop, econ JSON) pairs so it returns economy tables from a gameplay loop.

SFT · Structured Output
shot scripting-13

Camera Rig Plotter

Fine-tune Qwen/Qwen3.5-4B on 50 (scene, cam JSON) pairs so it outputs camera keyframes from a scene brief.

SFT · Structured Output
sound event maps-14

Audio Cue Ledger

Fine-tune Qwen/Qwen3.5-4B on 65 (event, sound JSON) pairs so it returns cue sheets from a gameplay event list.

SFT · Structured Output
extension points-15

Mod Hook Spindle

Fine-tune Qwen/Qwen3-8B on 70 (feature, mod JSON) pairs so it emits modding hook specs from a feature description.

SFT · Structured Output
puzzle layouts-16

Puzzle Grid Carver

Fine-tune Qwen/Qwen3.5-4B on 90 (rule, grid JSON) pairs so it outputs solvable puzzle grids from a rule set.

SFT · Structured Output
spatial anchors-17

XR Anchor Weaver

Fine-tune Qwen/Qwen3.5-9B on 60 (space, anchor JSON) pairs so it returns XR anchor graphs from a room brief.

SFT · Structured Output
animation graphs-18

Anim State Forge

Fine-tune Qwen/Qwen3-8B on 75 (char, anim JSON) pairs so it outputs state machines from a character move list.

SFT · Structured Output
tabletop conversion-19

Rulebook Tokenizer

Fine-tune Qwen/Qwen3.5-4B on 80 (rules, struct JSON) pairs so it emits digital rule structs from a tabletop text.

SFT · Structured Output
sequence beats-20

Cutscene Beat Coder

Fine-tune Qwen/Qwen3-8B on 55 (script, beats JSON) pairs so it returns shot beats from a cinematic script.

SFT · Structured Output
biome params-21

Terrain Rule Digger

Fine-tune Qwen/Qwen3.5-9B on 70 (bio, terra JSON) pairs so it outputs biome param sets from a terrain brief.

SFT · Structured Output
trophy graphs-22

Achievement Chain Smith

Fine-tune Qwen/Qwen3.5-4B on 65 (goal, ach JSON) pairs so it emits achievement trees from a design goal.

SFT · Structured Output
multiplayer state-23

Network Sync Slate

Fine-tune Qwen/Qwen3.5-9B on 60 (mode, sync JSON) pairs so it returns sync configs from a multiplayer mode.

SFT · Structured Output
riddle forms-24

Poetic Hint Carver

Fine-tune Qwen/Qwen3.5-4B on 50 (clue, poem JSON) pairs so it outputs structured riddles from a hint brief.

SFT · Structured Output
procedural level coherencer-0

Level Loop Forger

Fine-tune Qwen/Qwen3-8B on 60 level seeds so it emits tile graphs that pass a connectivity and difficulty grader.

RL · Programmatic Reward
narrative dialogue treesh-1

Dialogue Branch Smith

Fine-tune Qwen/Qwen3.5-4B on 40 scene specs so it writes branches that satisfy a consistency reward function.

RL · Programmatic Reward
GLSL heuristic shapingr-2

Shader Rule Carver

Fine-tune Qwen/Qwen3-8B on 50 effect prompts so generated shaders compile and pass a perf budget grader.

RL · Programmatic Reward
logic puzzle generationr-3

Puzzle Constraint Weaver

Fine-tune Qwen/Qwen3.5-4B on 70 puzzle templates so outputs pass a solvability checker reward.

RL · Programmatic Reward
resource economy curvesr-4

Economy Balance Tuner

Fine-tune Qwen/Qwen3-8B on 45 economy specs so it outputs curves passing a stability grader.

RL · Programmatic Reward
spatial navigation pathsr-5

XR Transit Mapper

Fine-tune Qwen/Qwen3.5-9B on 55 room graphs so it proposes paths passing a comfort grader.

RL · Programmatic Reward
combat encounter pacingr-6

Enemy Step Planner

Fine-tune Qwen/Qwen3-8B on 65 encounterbriefs so it emits beats passing a fairness grader.

RL · Programmatic Reward
accessible palette rulese-7

Color Palette Judge

Fine-tune Qwen/Qwen3.5-4B on 80 swatch sets so it returns palettes passing a contrast grader.

RL · Programmatic Reward
onboarding sequence trimr-8

Tutorial Step Pruner

Fine-tune Qwen/Qwen3.5-4B on 50 tutorials so it cuts steps passing a retention grader.

RL · Programmatic Reward
audio event placementr-9

Sound Cue Mapper

Fine-tune Qwen/Qwen3-8B on 60 timelines so it places cues passing a density grader.

RL · Programmatic Reward
quest dependency graphs-10

Quest Node Linker

Fine-tune Qwen/Qwen3-8B on 55 quest docs so it outputs DAGs passing a cycle grader.

RL · Programmatic Reward
keyframe rhythm fit-11

Animation Beat Aligner

Fine-tune Qwen/Qwen3.5-4B on 70 clips so it suggests keys passing a beat-error grader.

RL · Programmatic Reward
drop rate distributions-12

Loot Table Shaper

Fine-tune Qwen/Qwen3-8B on 45 tables so it emits rates passing a variance grader.

RL · Programmatic Reward
terrain zone mixing-13

Map Biome Blender

Fine-tune Qwen/Qwen3.5-9B on 50 heightmaps so it labels blends passing a slope grader.

RL · Programmatic Reward
control scheme feel-14

Input Buffer Coach

Fine-tune Qwen/Qwen3.5-4B on 65 schemes so it proposes buffers passing a misfire grader.

RL · Programmatic Reward
daily routine graphs-15

NPC Schedule Weaver

Fine-tune Qwen/Qwen3-8B on 60 NPCs so it builds routines passing a conflict grader.

RL · Programmatic Reward
multi-phase fight design-16

Boss Phase Smith

Fine-tune Qwen/Qwen3.5-9B on 40 bosses so it writes phases passing a lethality grader.

RL · Programmatic Reward
checkpoint granularity-17

Save State Pruner

Fine-tune Qwen/Qwen3.5-4B on 55 sessions so it picks points passing a loss-penalty grader.

RL · Programmatic Reward
VFX parameter bounds-18

Particle Rule Coder

Fine-tune Qwen/Qwen3-8B on 70 effects so it sets params passing an overload grader.

RL · Programmatic Reward
save-tree topology-19

Branching Save Forks

Fine-tune Qwen/Qwen3.5-4B on 50 trees so it forks states passing a depth grader.

RL · Programmatic Reward
camera shake limits-20

Camera Jolt Tamer

Fine-tune Qwen/Qwen3.5-4B on 65 shots so it tunes shake passing a nausea grader.

RL · Programmatic Reward
UI text compression-21

Tooltip Clarity Rater

Fine-tune Qwen/Qwen3.5-4B on 80 strings so it rewrites passing a readability grader.

RL · Programmatic Reward
agent movement laws-22

AI Steering Rule

Fine-tune Qwen/Qwen3-8B on 60 agents so it emits rules passing a collision grader.

RL · Programmatic Reward
changelog grouping-23

Patch Note Sorter

Fine-tune Qwen/Qwen3.5-4B on 55 logs so it groups items passing a category grader.

RL · Programmatic Reward
extension point linking-24

Mod Hook Mapper

Fine-tune Qwen/Qwen3-8B on 50 apis so it suggests hooks passing a safety grader.

RL · Programmatic Reward
studio tone consistencyr-0

House Voice Taster

Fine-tune Qwen/Qwen3.5-4B on 40 preference triples so any generation lands in the studio's house voice.

RL · Preference Pairs
gameplay rhythm tuninge-1

Level Pacing Palate

Fine-tune Qwen/Qwen3-8B on 60 preference triples so difficulty curves match the team's pacing philosophy.

RL · Preference Pairs
UX copy etiquettem-2

Death Screen Decorum

Fine-tune Qwen/Qwen3.5-4B on 50 preference triples so fail-state text stays respectful and on-brand.

RL · Preference Pairs
worldbuilding restraints-3

Lore Leak Lens

Fine-tune Qwen/Qwen3-8B on 70 preference triples so lore dumps avoid spoilers and stay subtle.

RL · Preference Pairs
onboarding voicer-4

Tutorial Tact Filter

Fine-tune Qwen/Qwen3.5-4B on 45 preference triples so tutorial lines feel guiding not nagging.

RL · Preference Pairs
scary suggestionr-5

Horror Hint Handler

Fine-tune Qwen/Qwen3-8B on 55 preference triples so horror text implies dread without explicit gore.

RL · Preference Pairs
palette tasten-6

Color Call Curation

Fine-tune Qwen/Qwen3.5-4B on 65 preference triples so palette notes match the art director's eye.

RL · Preference Pairs
antagonist voicey-7

Boss Banter Boundary

Fine-tune Qwen/Qwen3-8B on 50 preference triples so boss dialogue stays menacing not edgy.

RL · Preference Pairs
loot label tonet-8

Reward Text Restraint

Fine-tune Qwen/Qwen3.5-4B on 40 preference triples so item flavors avoid cheap hype wording.

RL · Preference Pairs
checkpoint moody-9

Save Scene Sensitivity

Fine-tune Qwen/Qwen3.5-4B on 48 preference triples so save prompts feel calm not intrusive.

RL · Preference Pairs
riddle framing-10

Puzzle Prompt Poise

Fine-tune Qwen/Qwen3-8B on 60 preference triples so puzzle hints tease without surrendering solution.

RL · Preference Pairs
immersive safety tone-11

XR Comfort Caution

Fine-tune Qwen/Qwen3.5-4B on 52 preference triples so XR warnings stay gentle yet clear.

RL · Preference Pairs
choice framing-12

Dialogue Branch Balance

Fine-tune Qwen/Qwen3-8B on 70 preference triples so dialogue options feel fair not loaded.

RL · Preference Pairs
cutscene tempo-13

Cinematic Cut Caution

Fine-tune Qwen/Qwen3.5-9B on 80 preference triples so cutscene direction notes avoid over-direction.

RL · Preference Pairs
inclusive copy-14

Accessibility Aside

Fine-tune Qwen/Qwen3.5-4B on 58 preference triples so helper text stays empowering not patronizing.

RL · Preference Pairs
secret framing-15

Easter Egg Etiquette

Fine-tune Qwen/Qwen3-8B on 44 preference triples so hidden-content notes stay clever not obscure.

RL · Preference Pairs
audio direction tone-16

Sound Cue Script Sense

Fine-tune Qwen/Qwen3.5-4B on 50 preference triples so SFX briefs stay evocative not technical.

RL · Preference Pairs
mod documentation-17

Mod Manifest Manners

Fine-tune Qwen/Qwen3.5-4B on 46 preference triples so mod docs stay welcoming not elitist.

RL · Preference Pairs
persona depth-18

Character Bio Bias

Fine-tune Qwen/Qwen3-8B on 62 preference triples so bios show depth not trope shorthand.

RL · Preference Pairs
marketing copy-19

Trailer Text Taste

Fine-tune Qwen/Qwen3.5-9B on 75 preference triples so trailer lines intrigue not oversell.

RL · Preference Pairs
playtest response-20

Feedback Flavor Filter

Fine-tune Qwen/Qwen3.5-4B on 54 preference triples so beta feedback stays constructive not harsh.

RL · Preference Pairs
environment storytelling-21

Setting Scroll Subtlety

Fine-tune Qwen/Qwen3-8B on 68 preference triples so scene description implies history not exposition.

RL · Preference Pairs
input guidance-22

Control Hint Courtesy

Fine-tune Qwen/Qwen3.5-4B on 42 preference triples so control tips stay succinct not cluttered.

RL · Preference Pairs
closure tone-23

Ending Echo Ethos

Fine-tune Qwen/Qwen3.5-9B on 78 preference triples so finale text resonates not preaches.

RL · Preference Pairs
update comms-24

Patch Note Poise

Fine-tune Qwen/Qwen3.5-4B on 56 preference triples so patch notes stay candid not defensive.

RL · Preference Pairs