File size: 1,094 Bytes
881f9f2 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 | # Mitigation rerun - Plan-Execute + Self-Ask with missing-evidence guard.
#
# Rerun this against the same frozen scenario set as the existing PE + Self-Ask
# baseline to populate results/metrics/mitigation_before_after.csv for #66.
EXPERIMENT_NAME="mitigation_missing_evidence_guard_pe_self_ask"
EXPERIMENT_CELL="Y"
EXPERIMENT_FAMILY="failure_mitigation"
SCENARIO_SET_NAME="smartgrid_multi_domain"
SCENARIOS_GLOB="data/scenarios/multi_*.json"
MODEL_ID="openai/Llama-3.1-8B-Instruct"
ORCHESTRATION="plan_execute"
MCP_MODE="baseline"
TRIALS=3
ENABLE_SELF_ASK=1
ENABLE_MISSING_EVIDENCE_GUARD=1
ENABLE_SMARTGRID_SERVERS=1
CONTRIBUTING_EXPERIMENTS="failure_mitigation,exp2_orchestration"
SCENARIO_DOMAIN_SCOPE="multi_domain"
MODEL_PROVIDER="vllm"
SERVING_STACK="insomnia_vllm"
QUANTIZATION_MODE="fp16"
MAX_MODEL_LEN=32768
TEMPERATURE=0.0
MAX_TOKENS=0
JUDGE_MODEL=""
LAUNCH_VLLM=1
VLLM_MODEL_PATH="models/Llama-3.1-8B-Instruct"
VLLM_PORT=8000
ENABLE_WANDB=1
WANDB_ENTITY="assetopsbench-smartgrid"
WANDB_PROJECT="assetopsbench-smartgrid"
WANDB_MODE="online"
TORCH_PROFILE="${TORCH_PROFILE:-1}"
|