You can deploy this model and start earning money today!

ModalityDance/
MRM-PRISM-V1

transformers

text-classification

HannahRoseKirk/prism-alignment

2601.18731

Skywork/Skywork-Reward-Llama-3.1-8B-v0.2

finetune

mit

endpoints_compatible

Views

No views yet

Meta Reward Modeling (MRM)

Overview

Meta Reward Modeling (MRM) is a personalized reward modeling framework designed to adapt to diverse user preferences with limited feedback.
Instead of learning a single global reward function, MRM treats each user as a separate learning task and applies a meta-learning approach to learn a shared initialization that enables fast, few-shot personalization.

MRM represents user-specific rewards as adaptive combinations over shared base reward functions and optimizes this structure through a bi-level meta-learning framework.
To improve robustness across heterogeneous users, MRM introduces a Robust Personalization Objective (RPO) that emphasizes hard-to-learn users during meta-training.

This repository provides trained checkpoints for reward modeling and user-level preference evaluation.

Evaluation

The model is evaluated using user-level preference accuracy with few-shot personalization.
Inference follows the same adaptation procedure used during training: for each user, the reward weights are initialized from the meta-learned initialization and updated with a small number of gradient steps on user-specific preference data.

Example evaluation script

bash

python inference.py \
  --embed_pt data/emb/prism/V1.pt \
  --meta_json data/emb/prism/V1.json \
  --ckpt path/to/checkpoint.pt \
  --dataset PRISM \
  --seen_train_limit -1 \
  --unseen_train_limit -1 \
  --hidden_layers 2 \
  --inner_lr 1e-3 \
  --eval_inner_epochs 1 \
  --val_ratio 0.9 \
  --score_threshold -1 \
  --seed 42 \
  --device cuda:0

Usage Example

This example shows a typical workflow for a single user:

encode text pairs with Skywork/Skywork-Reward-Llama-3.1-8B-v0.2 into embeddings,
adapt the MRM on the user's few-shot examples (update shared_weight only),
run inference on new pairs for that same user.

python

import torch
from copy import deepcopy
from transformers import AutoTokenizer, AutoModelForSequenceClassification

from utils import bt_loss
from train import MRM
from inference import load_ckpt_into_model


@torch.no_grad()
def encode_pairs(model, tokenizer, pairs, device="cuda"):
    model.eval()
    ch, rj = [], []
    for ex in pairs:
        conv = ex["prompt"]
        for key, buf in [("chosen", ch), ("rejected", rj)]:
            ids = tokenizer.apply_chat_template(
                conv + [{"role": "assistant", "content": ex[key]}],
                tokenize=True, return_tensors="pt"
            ).to(device)
            out = model(ids, output_hidden_states=True)
            buf.append(out.hidden_states[-1][0, -1].float().cpu())
    return torch.stack(ch), torch.stack(rj)


def adapt_single_user(base_model, support_ch, support_rj, inner_lr=1e-3, inner_epochs=5, device="cuda"):
    model = deepcopy(base_model).to(device).train()
    opt = torch.optim.Adam([model.shared_weight], lr=inner_lr)
    support_ch, support_rj = support_ch.to(device), support_rj.to(device)
    for _ in range(inner_epochs):
        opt.zero_grad()
        loss = bt_loss(model(support_ch), model(support_rj))
        loss.backward()
        opt.step()
    return model.eval()


@torch.no_grad()
def infer_on_pairs(model, ch, rj, device="cuda"):
    return model(ch.to(device)), model(rj.to(device))


device = "cuda" if torch.cuda.is_available() else "cpu"

MODEL_PATH = "Skywork/Skywork-Reward-Llama-3.1-8B-v0.2"
tokenizer = AutoTokenizer.from_pretrained(MODEL_PATH)
llm = AutoModelForSequenceClassification.from_pretrained(
    MODEL_PATH, num_labels=1, torch_dtype=torch.bfloat16, device_map=device
)

CKPT_PATH = "ckpt/model.pt"
mrm = MRM(in_dim=4096, hidden_sizes=[2], use_bias=False)
load_ckpt_into_model(mrm, CKPT_PATH, device)

support_pairs = [
    {
        "prompt": [{"role": "user", "content": "TL;DR this post: I tried waking up at 5am for a month and tracked my productivity."}],
        "chosen": "Waking up early helped at first, but long-term productivity depended more on sleep quality than wake-up time.",
        "rejected": "The post is about waking up early and productivity.",
    },
    {
        "prompt": [{"role": "user", "content": "Summarize the main point: I switched from iPhone to Android after 10 years."}],
        "chosen": "The author values customization and battery life more than ecosystem lock-in, which motivated the switch.",
        "rejected": "The author bought a new phone.",
    },
]

sup_ch, sup_rj = encode_pairs(llm, tokenizer, support_pairs, device)
user_mrm = adapt_single_user(mrm, sup_ch, sup_rj, device=device)

test_pairs = [
    {
        "prompt": [{"role": "user", "content": "TL;DR: I quit my job to freelance and here is what I learned in 6 months."}],
        "chosen": "Freelancing offers flexibility but requires strong self-discipline and financial planning to be sustainable.",
        "rejected": "The author talks about quitting a job and freelancing.",
    }
]

test_ch, test_rj = encode_pairs(llm, tokenizer, test_pairs, device)
s_ch, s_rj = infer_on_pairs(user_mrm, test_ch, test_rj, device)

print("reward(chosen)  =", s_ch.tolist())
print("reward(rejected)=", s_rj.tolist())

Citation

If you use this model or code in your research, please cite:

bibtex

@inproceedings{cai2026MRM,
  title={One Adapts to Any: Meta Reward Modeling for Personalized LLM Alignment},
  author={Hongru Cai and Yongqi Li and Tiezheng Yu and Fengbin Zhu and Wenjie Wang and Fuli Feng and Wenjie Li},
  booktitle={Proceedings of the 49th International ACM SIGIR Conference on Research and Development in Information Retrieval},
  series={SIGIR '26},
  year={2026}
}

License

This model is released under the MIT License.

ModalityDance/MRM-PRISM-V1