Spaces:
Sleeping
Sleeping
| name: permanence | |
| version: 1.1.0 | |
| spec_version: "0.1" | |
| entry_point: permanence.openenv_env:PermanenceOpenEnv | |
| description: > | |
| A reversibility-aware RL environment that trains LLMs to predict | |
| whether their actions are recoverable before taking them. Built on | |
| three operational-semantics simulators (filesystem, git, database) | |
| where R-levels (R1-R5) are derived from world state at execution | |
| time — not hardcoded tags. Agents must emit a <reversibility/> tag | |
| alongside each <action/>, and are scored on both task completion | |
| and prediction calibration. | |
| author: chanikya | |
| email: chanikyac01@gmail.com | |
| huggingface_repo: chane35/permanence | |
| tags: | |
| - openenv | |
| - world-modeling | |
| - long-horizon-planning | |
| - reinforcement-learning | |
| - agent-safety | |
| type: chat | |
| app: | |
| module: server.app | |
| object: app | |
| port: 7860 | |
| themes: | |
| primary: world_modeling | |
| secondary: | |
| - long_horizon_planning | |
| tasks: | |
| - id: task_correction | |
| difficulty: 1 | |
| description: Report error correction with irreversible external communication risk | |
| score_range: [0.0, 1.0] | |
| - id: task_conflict | |
| difficulty: 2 | |
| description: Personnel conflict resolution with irreversible HR action risk | |
| score_range: [0.0, 1.0] | |
| - id: task_launch | |
| difficulty: 3 | |
| description: Product launch decision with irreversible public commitment risk | |
| score_range: [0.0, 1.0] | |
| - id: task_crisis | |
| difficulty: 4 | |
| description: Crisis response requiring mandatory irreversible action under time pressure | |
| score_range: [0.0, 1.0] | |
| - id: task_cascade | |
| difficulty: 5 | |
| description: Multi-step resolution where premature action permanently locks all downstream steps | |
| score_range: [0.0, 1.0] | |
| - id: task_db_migration | |
| difficulty: 5 | |
| description: Production schema migration with cascade mechanic grounded in real infrastructure operations (backup, dry-run, maintenance window, DDL apply, rollback, backfill, column drop) | |
| score_range: [0.0, 1.0] | |
| environment: | |
| observation_type: text | |
| action_type: text | |
| multi_agent: false | |
| persistent_within_episode_state: true | |
| max_observation_tokens: 1800 | |
| reward_range: [-0.5, 1.0] | |
| max_steps_per_episode: 15 | |
| reward_components: | |
| task_completion: 0.40 | |
| prediction_accuracy: 0.30 | |
| option_preservation: 0.20 | |
| catastrophe_penalty: 0.10 | |
| training: | |
| recommended_model: meta-llama/Llama-3.2-3B-Instruct | |
| recommended_algorithm: grpo | |
| recommended_framework: unsloth | |
| episodes: 1500 | |
| warmup_sft_episodes: 20 | |
| gpu_hours: 7 | |
| cost_usd: 20 | |
| novelty: | |
| - Within-episode persistent world state — no prior OpenEnv environment has this | |
| - R-level computed from world state at runtime, not static tag | |
| - Prediction accuracy as first-class reward component | |
| - Symmetric penalty on misclassification — over-caution punished equally to under-caution | |
| - Task 4 requires taking irreversible action correctly — proves no caution training | |