Downloads · 30 days
0
core16-harmless-0825-1358
agentic-moral-alignment/core16-harmless-0825-1358
core16-harmless-0825-1358 is a machine learning model from agentic-moral-alignment. Use it for the machine learning task on the model card, and read the license before you ship it in a product. It is set up for peft.
LoRA adapters from mt-moral-alignment run core16-harmless-0825-1358, one subfolder per checkpoint, converted from verl FSDP shards. Per-episode logs: https://huggingface.co/datasets/agentic-moral-alignment/mtma/tree/m…
Access
Public
Updated Aug 26, 2026
Repo size
1 GB
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors1 GB · 100%
At a glance
- Library
- peft
- Access
- Public
- Created
- Aug 26, 2026
- Updated
- Aug 26, 2026
- SHA
- 30be0514
Base models
Model card
From the Hugging Face model README
core16-harmless-0825-1358
LoRA adapters from mt-moral-alignment run core16-harmless-0825-1358, one subfolder per checkpoint,
converted from verl FSDP shards. Per-episode logs:
https://huggingface.co/datasets/agentic-moral-alignment/mtma/tree/main/runs/train/core16-harmless-0825-1358
| step | load |
|---|---|
| 5 | subfolder="step_5" |
| 10 | subfolder="step_10" |
| 15 | subfolder="step_15" |
| 20 | subfolder="step_20" |
| 25 | subfolder="step_25" |
| 30 | subfolder="step_30" |
| 35 | subfolder="step_35" |
| 40 | subfolder="step_40" |
base = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3.5-4B-Base")
model = PeftModel.from_pretrained(base, "agentic-moral-alignment/core16-harmless-0825-1358", subfolder="step_40")
WARNING: vLLM cannot serve these adapters on Qwen3.5 -- hybrid-GDN LoRA is a
silent no-op (vllm#49354). Merge first: model.merge_and_unload().save_pretrained(...).
Frozen run config
# frozen for core16-harmless-0825-1358 (2026-08-25T13:58:58Z, git 4895099)
STAGE: train
MODEL: Qwen/Qwen3.5-4B-Base
GPUS: 2
TP: 1
MAX_SEQS: 384
GPU_MEM_UTIL: 0.9
GAMES: core16
BATCH_GAME: 4
START_SAMPLE: True
START_RESAMPLE_STEP: True
TURN_CAP: 64
N: 8
STEPS: 40
LR: 2e-05
TOP_P: 0.95
ROLLOUT_SEED: 42
FRAMING: game-goals
LAM: 1
TASK_WEIGHT: 0
MORAL: deon_soft_harmless
RETURN_FN: reward_to_go
BASELINE_FN: state_mean
NORMALIZE_FN: none
GAMMA: 0.0
OMEGA: 0
TD_LAMBDA: 0.8
THINKING: True
MAX_TURN_TOKENS: 4096
ENV_TIMEOUT_S: 60
AGENT_WORKERS: 16
REWARD_WORKERS: 1
BYPASS: True
GRAD_CKPT: True
CALC_ENTROPY: True
SAVE_FREQ: 5
NOSET: 1
PUSH: 1
PROMPT_LEN: 4096
RESPONSE_LEN: 65536
MAX_MODEL_LEN: 69632
TOKEN_LEN_PER_GPU: 139264
LORA_RANK: 32
LORA_ALPHA: 64
- Library
- peft
- Created
- Aug 26, 2026
- Updated
- Aug 26, 2026