Spaces:
Sleeping
Sleeping
Commit History
-
Add Dockerfile
fa2121f -
readme updated
6d8dd98 -
chore: add .gitattributes for git-lfs tracking of mp4 and joblib
a74c791 -
mp4 removed
02948b2 -
mp4 replaced
5b33916 -
final mvp created
98b952a -
mvp version 2 created
775ccbd -
feat(phase12): RetentionCurveSimulator, R10, 150-sample dataset, model trained, 14 tests PASS, gate PASS
79cb04a -
phase 11 passed
09f7d63 -
feat(phase10): ABScriptEnv, ContrastiveReward, A/B rollout fn, 25 tests PASS, gate PASS
cfe83fc -
feat(phase9): platform reward divergence — PlatformRegistry, R9 PlatformPacing, R1/R2 platform-aware, 20 tests PASS, gate PASS
998d987
-
Phase 8 implemented
dfa9070 -
Phase 7 implemented
0e4f105 -
Phase 6 implemented
5c28dc0 -
Phase 5 implemented
e6b6793 -
feat(phase4): critic escalation engine, difficulty tracker, env wiring, gate PASS
ebae6ab -
phase2 implemented
258783b -
phase1 implememted
0290d7a -
feat(phase1): OpenEnv scaffold + R1/R2 rewards — PHASE 1 GATE PASS
- environment/actions.py: ActionType enum + ArbitratorAction model - environment/observations.py: RewardComponents (weighted, normalised), DebateRound, Observation - environment/episode_state.py: mutable episode state dataclass - environment/env.py: ViralScriptEnv — Gymnasium-compatible reset/step/state, difficulty tiers, anti-gaming wired - rewards/r1_hook_strength.py: 5-check rule-based hook scorer (promise, curiosity, specificity, front-load, anti-filler) - rewards/r2_coherence.py: sentence-transformers cosine similarity with 4-range score mapping + embedding cache - rewards/reward_aggregator.py: catastrophic-drop (>0.2) + action-diversity anti-gaming rules - agents/rewriter.py: RewriterAgent wrapping LLMBackend with unified diff output - scripts/run_dummy_episode.py: demo runner with rich output and gate check - tests/test_rewards.py: 10 tests — R1 (5), R2 (2), aggregator (3) - tests/test_environment.py: 6 tests — all LLM calls mocked via golden fixtures Co-Authored-By: Claude Son…41ea373 -
feat(phase0): add Groq backend, set as default (llama-3.3-70b-versatile)
5fcd0ee -
feat(phase0): switch default model to Qwen2.5-1.5B-Instruct
bb30917 -
refactor(phase0): replace Anthropic SDK with LLMBackend abstraction (default Qwen)
639b641 -
test(phase0): add full mocked test suite for CriticAgent
5a386b9 -
feat(phase0): add run_critic_gate CLI with rich output and fixture saving
7414383 -
feat(phase0): add CriticEvaluator with rule-based gate logic
fea7356 -
feat(phase0): add 10 test scripts with known flaws
4e6f033 -
feat(phase0): add CriticAgent, CritiqueClaim, CritiqueOutput models
5ccdeee -
feat(phase0): scaffold directories and requirements
806216c -
base structure of the project formed
b6e19c7