A student scorer from MIRA (Mid-training Rubric Anchoring for Source-Aware Data Selection), fine-tuned to score Chinese HTML web-design documents along a group-specific set of anchor rubric dimensions.

📄 Paper: MIRA: Mid-training Rubric Anchoring for Source-Aware Data Selection (EMNLP 2026) 💻 Code: https://github.com/Multilingual-Multimodal-NLP/mira

TL;DR

MIRA is a source-aware data selection framework for heterogeneous mid-training corpora. Instead of applying a single global quality rubric, MIRA (1) clusters sources into capability-coherent groups, (2) lets a frontier teacher (Kimi-K2.6) freely propose rubric dimensions and anchors them per group, (3) distills the anchored teacher into a lightweight per-group student scorer, and (4) applies reliability-aware aggregation with per-source retention thresholds.

This repository is one of those student scorers — variant 1 in the Text family, specialized for Chinese HTML web-design documents. Given an in-distribution record, it produces a numerical score and a short rationale for every anchor dimension in this group's rubric.

Model summary


Architecture	Mixture-of-Experts decoder (35B total / ≈3B active params)
Base model	Qwen3.5-35B-A3B-Base
Fine-tuning	Full-parameter SFT on Kimi-K2.6 anchored teacher labels
Domain	Chinese-dominant flat-text documents describing front-end / HTML implementations — the `html_cpt` source (score_data_5B + PT_DATA_TRAIN, median ~8.5K chars per doc).
Anchor rubric	15 group-specific dimensions (`group_B_dim_anchors.jsonl` in the project repo)
Source count	1 text source
Output	Structured (score, rationale) per anchor dimension
Precision	BF16
License	Apache-2.0 (inherits from Qwen3)

Sources covered

This scorer is calibrated for the following mid-training sources in the Text / Chinese HTML web-design group:

Source	Description
`html_cpt`	Chinese HTML design / CPT documents (~8.5K chars median)

The full source-grouping report (KMeans k=4 / 5 clusters, intra-group cosine similarities) is in the project repo.

Anchor dimensions (15 slots)

The scoring rubric for this group, discovered via Kimi-K2.6 free-form judging and clustered into 15 anchor dimensions (KMeans k=15 over the group's dim-score embeddings). Dimensions below are sorted by cluster size — larger clusters dominate the corpus and carry more signal. Anchor names are read verbatim from this group's group_B_dim_anchors.jsonl; some names recur across slots because semantically related but distinct rubric facets were clustered separately by the teacher.

Slot	Dimension	Cluster size
A1	Chinese Description Quality	16,842
A2	Feature Completeness	14,072
A3	Frontend Technical Appropriateness	13,949
A4	JavaScript Interactivity Quality	11,308
A5	Semantic HTML Structure	10,790
A6	Code Readability & Organization	8,645
A7	Accessibility (a11y) Compliance	8,303
A8	Visual Design & Polish	8,142
A9	Practical Real-world Utility	7,758
A10	Performance Awareness	6,812
A11	Technical Correctness	6,298
A12	Security & Safety	6,245
A13	Responsive Design Implementation	6,184
A14	Responsive Design Execution	3,647
A15	Accessibility Compliance	560

The scorer outputs one [Ai] <dimension>: <score>/10 — <rationale> line per slot, plus overall, training_recommendation, domain_tag, and brief.

Where this model fits in the MIRA pipeline

┌──────────────────┐  ┌──────────────────┐  ┌──────────────────┐  ┌──────────────────┐
│ 1. Rubric        │  │ 2. Anchored      │  │ 3. Reliability   │  │ 4. Data          │
│    Discovery     │→ │    Judge         │→ │    Aggregation   │→ │    Selection     │
│ (Kimi-K2.6,      │  │    Distillation  │  │ (mask unreliable │  │ (per-source      │
│  free-form       │  │ ◀── THIS MODEL   │  │  src×dim cells)  │  │  retention)      │
│  judging)        │  │                  │  │                  │  │                  │
└──────────────────┘  └──────────────────┘  └──────────────────┘  └──────────────────┘

MIRA-Text-Group1 lives in Stage 2: it scores the full Text / Chinese HTML web-design corpus so that downstream stages can apply reliability masking and source-aware retention.

Intended use

Primary: Score Chinese HTML web-design documents on this group's anchor dimensions to drive source-aware data selection and filtering.
Secondary: Research on rubric distillation, semantic quality scoring, and reliability diagnostics for heterogeneous training corpora.

Not intended for:

General-purpose chat or instruction following — fine-tuned to emit structured scores, not freeform dialogue.
Single-shot quality judgments without the anchor-dimension prompt template — outputs will be miscalibrated.
Records outside the Text / Chinese HTML web-design group; use the matching sibling scorer instead.

Deployment

The scorer is designed to be served via vLLM behind an OpenAI-compatible endpoint and called in batch from the MIRA scoring pipeline.

1. Serve with vLLM (recommended)

bash

vllm serve whw06/MIRA-Text-Group1 \
    --tensor-parallel-size 8 \
    --dtype bfloat16 \
    --max-model-len 65536 \
    --max-num-batched-tokens 131072 \
    --gpu-memory-utilization 0.9 \
    --trust-remote-code \
    --port 8000

Why these values (verified on H200 141GB during the paper's per-source evaluation):

max-model-len=65536 — 2× the mid-training cutoff. Records can hit ~60K tokens for densely-tokenized sources; 40K runs into prompt-overflow errors.
max-num-batched-tokens=131072 — supports two full-length sequences per scheduling step.
gpu-memory-utilization=0.9 — 35B BF16 weights take ~70GB, leaving ~57GB KV cache. Roughly 4 concurrent 65K-context sequences per GPU.
8-way tensor parallel works well for the 35B MoE on a single 8×H200/A100 node.

2. Call from Python

python

from openai import OpenAI

client = OpenAI(base_url="http://localhost:8000/v1", api_key="EMPTY")

resp = client.chat.completions.create(
    model="whw06/MIRA-Text-Group1",
    messages=[
        {"role": "system", "content": SYSTEM_PROMPT},   # group-B anchor calibration
        {"role": "user",   "content": USER_PROMPT},     # record + [A1]..[A15] template
    ],
    temperature=0.7,
    top_p=0.95,
    max_tokens=2048,
)
print(resp.choices[0].message.content)

3. Prompt template

The user message asks for one structured line per anchor dimension (top-15 of this group):

[A1] {anchor_dim_1}: <score>/10 — <justification>
[A2] {anchor_dim_2}: <score>/10 — <justification>
...
[A15] {anchor_dim_15}: <score>/10 — <justification>
overall: <0-100>
training_recommendation: <keep | downsample | drop>
domain_tag: <short tag>
brief: <one-sentence summary>

The system prompt embeds the top-12 anchor calibration references (canonical examples from clustering) so the student matches the teacher's scoring scale. The full prompt builder, anchor JSONL files, and output parser are in the project repo's scoring/score_text_anchored.py.

Training details


Teacher	Kimi-K2.6 (free-form rubric discovery in Phase 1; anchored re-scoring in Phase 2)
Training data	Kimi-K2.6 anchored labels on this group's Phase-2 corpus, split into a distillation set + a held-out validation split for reliability diagnostics
Loss	Standard next-token CE over (score, rationale) labels for every anchor dimension
Hyperparameters	Held constant across all MIRA student scorers; full settings in paper Appendix A.4
Validation	Per-dimension teacher–student MAE and Spearman ρ on a held-out split; dimensions failing reliability thresholds are masked post-hoc (Figure 3 in the paper)

Training loss / step curve is preserved in trainer_state.json for full reproducibility.

Headline results (from the paper)

End-to-end downstream evaluation: Qwen2.5-Coder-14B mid-trained on 25B-token MIRA-selected subsets vs. baselines, then SFT, evaluated on 9 code benchmarks across 4 categories.

Method	Code Gen	MultiplE	SQL (EX)	SWE-Multi	Macro Avg
Base + SFT (no mid)	53.91	72.57	64.24	3.67	48.60
Raw Mixture (50B)	53.71	67.42	94.18	40.00	63.83
Random (25B)	52.71	71.44	91.03	35.00	63.23
DataMan (25B)	53.82	71.38	93.84	33.00	63.01
DSIR (25B)	48.74	67.26	95.20	27.00	59.55
PPL (25B)	50.52	57.74	90.66	20.00	54.73
MIRA-Global (25B)	53.12	67.84	94.26	32.00	61.81
MIRA-Group (25B)	54.53	71.85	94.08	36.33	64.20
MIRA-Source (25B)	54.18	72.84	94.38	30.33	62.93

MIRA-Group matches the full 50B-token raw mixture while using only half the tokens, and out-performs all 25B-token selection baselines on the macro average. This scorer is one of the 12 student models used by the MIRA-Group variant.

Sibling models

MIRA releases one student scorer per source-group variant. Use the matching scorer for each record's format:

Agent: whw06/MIRA-Agent-Group1 · -Group2 · -Group3 · -Group4
QA: whw06/MIRA-QA-Group1 · -Group2 · -Group3 · -Group4 · -Group5
Text: MIRA-Text-Group1 (this model) · -Group2 · -Group3

Limitations

MIRA addresses source-aware filtering only. Source discovery, mixture-ratio design, curriculum scheduling, deduplication and contamination control remain orthogonal concerns.
This scorer is calibrated against the Text / Chinese HTML web-design group; cross-domain transfer is not advised — use the matching sibling for other source formats.
Some anchor dimensions exhibit high teacher–student MAE and are masked post-hoc during aggregation (see paper §3.4). The model still emits scores for masked dimensions; downstream consumers should re-apply the reliability mask from the project repository.
Calibrated on 1 source within this group; behavior on out-of-distribution formats is unverified.

Citation

bibtex

@inproceedings{wang2026mira,
  title     = {MIRA: Mid-training Rubric Anchoring for Source-Aware Data Selection},
  author    = {Wang, Haowen and Du, Yaxin and Yang, Jian and Wu, Jiajun and
               Liu, Shukai and Zhang, Yuxuan and Wang, Pingjie and Chen, Siheng and
               Zheng, Tuney and Zhou, Ming and Liu, Xianglong},
  booktitle = {Proceedings of the 2026 Conference on Empirical Methods in Natural Language Processing (EMNLP)},
  year      = {2026}
}

Acknowledgments

Built on Qwen3.5-35B-A3B-Base and the Megatron-LM training stack. Teacher labels generated with Kimi-K2.6.