---
title: 'RedOne: Dual Systems in Radiology & SNS'
url: https://www.emergentmind.com/topics/redone
type: topic
---

# RedOne: Dual Systems in Radiology & SNS

Searching arXiv for the provided RedOne-related papers to ground the article in the latest records.
RedOne is a polysemous label in recent machine-learning literature. In one line of work, it denotes **CXR-ReDonE**, a chest X-ray radiology report generation system whose central aim is to remove hallucinated references to non-existent prior examinations from generated impressions [2210.06340]. In another, it denotes a **family of domain-specific large language models for Social Networking Services (SNS)** developed by Xiaohongshu, first as **RedOne** and later as **RedOne 2.0**, with post-training recipes designed for heterogeneous SNS tasks, bilingual content, and robustness under distribution shift [2507.10605] [2511.07070]. The shared name does not imply a shared technical lineage; the cited works concern separate application domains, model classes, datasets, and evaluation protocols.

## 1. Terminological scope and disambiguation

In the cited literature, “RedOne” refers to two unrelated systems.

| Usage of “RedOne” | Domain | Core characterization |
|---|---|---|
| **CXR-ReDonE** | Radiology report generation | “Contrastive X-Ray Report Determination Employing Prior Reference Removal” |
| **RedOne** | SNS-domain LLM | Xiaohongshu’s original SNS-oriented LLM |
| **RedOne 2.0** | SNS-domain LLM | RL-prioritized successor to the original RedOne |

For radiology, the shorthand “RedOne” arises informally from **CXR-ReDonE** (“Re-Don-E → RedOne”), while the paper explicitly introduces **CXR-ReDonE** as the main new system [2210.06340]. For SNS, **RedOne** is the formal model name, and **RedOne 2.0** is a redesigned successor that replaces a classical SFT-dominated recipe with a progressive, RL-prioritized post-training paradigm [2507.10605] [2511.07070].

This dual usage is important because the two research programs address different failure modes. CXR-ReDonE targets **hallucinated temporal comparison language** in medical reporting. RedOne and RedOne 2.0 target **domain adaptation for social-network workloads**, including moderation, retrieval, dialogue, user-behavior modeling, and SNS-style translation.

## 2. CXR-ReDonE: RedOne in chest radiography

CXR-ReDonE is a radiology report generation system for chest X-rays that is designed specifically to avoid hallucinated references to non-existent prior exams [2210.06340]. Its task is standard report generation from a chest X-ray image or images, producing the free-text radiology report’s **impression** section. The system is built by retraining **CXR-RePaiR**, a retrieval-based report generator, on a cleaned version of **MIMIC-CXR** in which references to priors are removed or rewritten; this adapted dataset is called **MIMIC-PRO**.

The motivating problem is that radiology reports in routine practice often refer to **priors**, meaning previous imaging studies and their associated reports for the same patient. In the source data, phrases such as “Compared to the prior radiograph,” “No interval change from prior study,” “Heart size is stable,” and “Again seen is” encode comparative information over time rather than pure description of the current study. In **MIMIC-CXR**, **173,822 of 226,759 reports (76.3%)** contain at least one of 18 “prior-like” keywords, including *change*, *unchanged*, *prior*, *stable*, *interval*, *previous*, and *again* [2210.06340].

The paper’s core claim is that models trained on such text internalize the statistical regularity that good radiology reports often contain prior-related language. At inference time, however, deployed systems frequently do **not** have access to previous exams or reports. The result is a clinically important hallucination mode: models generate statements such as “Compared to the prior study…” or “Heart size is unchanged from the prior radiograph” even when only a single current X-ray is available. The paper shows that this failure mode occurs in systems such as **CXR-RePaiR**, **R2Gen**, and **M² Trans**, and argues that such hallucinations misrepresent the evidential basis of the report and can undermine clinical trust [2210.06340].

The dataset context is equally central. **MIMIC-CXR** contains **377,110 images** from **227,835 studies**, and the paper uses a curated subset of **226,759 reports**, referred to in the paper simply as “MIMIC-CXR.” Reports are semi-structured, typically including **Findings** and **Impression**, and the benchmark role of MIMIC-CXR in chest X-ray report generation makes prior-reference contamination a systemic issue rather than a narrow artifact of one model.

## 3. Prior-reference removal and construction of MIMIC-PRO

The central contribution behind CXR-ReDonE is a **data-centric pipeline** for removing prior-related language from the training corpus [2210.06340]. The paper proposes two methods: a **GPT-3-based few-shot rewriting pipeline** named **FilBERT+GPT-3**, and a **BioBERT-based token classification pipeline** named **GILBERT**. Both are developed using a manually curated shared corpus of **103 report pairs**, where each pair contains an original report and a ground-truth edited version with prior references removed or rewritten.

The first method, **FilBERT+GPT-3**, is sentence-centric. **FilBERT** is a BioBERT base model with a sequence classification head that predicts whether a sentence contains any reference to a prior. It is initialized from `dmis-lab/biobert-base-cased-v1.2`, trained for **10 epochs** with Adam at learning rate \(2 \times 10^{-5}\) and \(\epsilon = 10^{-8}\), with batch size **16** on one Tesla P100 GPU. On the held-out test set it achieves **accuracy \(= 0.907\)**, with **false positive rate \(= 5.55\%\)** and **false negative rate \(= 3.70\%\)** at threshold \(0.5\) [2210.06340]. This filter is used because running GPT-3 on every report sentence in MIMIC-CXR is estimated to cost **\$92,000**, whereas filtering reduces the estimate to **\$16,560**.

Flagged sentences are then rewritten by **GPT-3 (DaVinci)** using few-shot prompting. The paper reports that **DaVinci** outperforms Curie, Babbage, and Ada, with **\(F_1 = 0.5569\)** versus **0.37**, **0.33**, and **0.23**. A **temperature sweep** finds the best result at **temperature \(= 0.3\)**, and experiments over subreports of \(n \in \{1,2,3,4\}\) sentences show that **\(n = 1\)** performs best, indicating that GPT-3 is more reliable as a sentence-level rewriter than as a whole-report editor [2210.06340].

The second method, **GILBERT**, formulates prior removal as a token-classification problem. Each token is labeled as either **KEEP** or **REMOVE**, rather than with BIO tags. Because BioBERT uses WordPiece tokenization, the tokenizer is modified so that every subword receives a label. The model uses BioBERT with a token-classification head and standard token-level cross-entropy,
$$
\mathcal{L} = - \sum_{i=1}^{T} \left[ y_i \log p_i(\text{REMOVE}) + (1 - y_i) \log p_i(\text{KEEP}) \right].
$$
Training is implemented in PyTorch for **10 epochs**, each with **100 steps**, with training batch size **4**, test batch size **2**, gradient clip norm **10**, Adam learning rate \(1 \times 10^{-5}\), on a single Tesla T4 GPU [2210.06340].

On the shared test corpus, **GILBERT achieves \(F_1 = 0.84\)**, substantially higher than **FilBERT+GPT-3**, which achieves **\(F_1 = 0.56\)**. When GILBERT is applied to all **226,759 reports** in MIMIC-CXR, the total number of prior keyword instances drops from **259,376** to **82,074**, a **>68.3% reduction**. Because of this higher \(F_1\), lower cost, and better scalability, GILBERT is selected as the primary method for constructing **MIMIC-PRO** [2210.06340].

The paper also reports a manually adapted evaluation set: from the standardized MIMIC-CXR test set of **2,192 images**, a board-certified radiologist and two senior medical students edit reports to remove prior phrases or rewrite them as current-state descriptions. This expert-edited set serves as the ground-truth test set for evaluating report-generation systems that are not supposed to reference priors.

## 4. Retrieval architecture, evaluation, and limitations of CXR-ReDonE

CXR-ReDonE inherits the retrieval formulation of **CXR-RePaiR** but changes the training data and the vision-language backbone [2210.06340]. CXR-RePaiR treats report generation as retrieval from a corpus \(\mathcal{R} = \{r_1, \dots, r_n\}\), encoding reports with a text encoder \(g(\cdot)\) and images with an image encoder \(h(\cdot)\), then scoring image-report compatibility by dot product:
$$
f(r, x) = g(r) \cdot h(x).
$$
Prediction can operate at the **report level**, returning the report \(r\) that maximizes \(f(r,x)\), or at the **sentence level**, composing a synthetic report from the top \(k\) retrieved sentences.

CXR-ReDonE retrains this framework on **MIMIC-PRO** and replaces the original **CLIP** backbone with **ALBEF** (“Align Before Fuse”), while preserving the core retrieval logic. The paper emphasizes that the main innovation is not a new generator architecture but a combination of **prior-cleaned training text** and a stronger image-text aligner. Training uses **60 epochs**, **100 steps per epoch**, on **4 Quadro RTX-8000 GPUs** [2210.06340].

Evaluation uses **BERTScore**, **\(s_{emb}\)** based on **CheXbert embedding similarity**, and **RadGraph \(F_1\)**, rather than BLEU, METEOR, or CIDEr, which the paper argues are not well suited to radiology. On the expert-edited test set, CXR-ReDonE improves over the baseline trained on original MIMIC-CXR in every evaluated configuration. At **report-level retrieval**, the baseline scores **BERTScore \(= 0.2083\)**, **\(s_{emb} = 0.3410\)**, and **RadGraph \(F_1 = 0.0895\)**, whereas CXR-ReDonE scores **0.2160**, **0.3601**, and **0.0925**. At sentence-level retrieval with **\(k = 2\)**, where CXR-ReDonE achieves its best BERTScore, the baseline scores **0.2292 / 0.3822 / 0.1045**, whereas CXR-ReDonE reaches **BERTScore \(= 0.2351\)**, **\(s_{emb} = 0.3859\)**, and **RadGraph \(F_1 = 0.1056\)**, corresponding to a **2.57% absolute improvement** in BERTScore [2210.06340]. At **\(k = 3\)**, CXR-ReDonE achieves the best reported **RadGraph \(F_1 = 0.1112\)**, a **+2.68%** improvement over the baseline.

Qualitative examples show the same pattern. Systems trained on the original corpus often generate “stable,” “again,” “unchanged,” or “decreased since previous,” whereas CXR-ReDonE tends to produce descriptions of current device positioning and lung findings without unsubstantiated comparative language. The paper interprets this as evidence that removing prior references from training data improves clinical content fidelity rather than degrading it [2210.06340].

The limitations are explicit. **Residual hallucinations** remain because prior keyword counts are reduced by about **68%**, not eliminated. **Grammar and fluency** can degrade because GILBERT deletes tokens without grammatical repair; the paper mentions a future grammatical error correction layer such as **GECToR**. **Temporal information** is also genuinely useful in practice, so removing all prior references is appropriate only for the setting where priors are unavailable. Additional constraints include dependence on BioBERT and GPT-3 behavior, restriction to chest radiography and MIMIC-CXR, and the absence of a prospective user study or radiologist satisfaction metric.

## 5. RedOne as a domain-specific LLM for Social Networking Services

In the SNS literature, **RedOne** is a domain-specific LLM developed by Xiaohongshu for Social Networking Services, rather than a radiology system [2507.10605]. The motivation is that many existing SNS applications rely on separate single-task finetuned models for sentiment analysis, rumor detection, content moderation, search relevance, or other scenarios. The paper argues that such specialization causes performance plateaus, fragmented deployment across scenarios, and a mismatch with SNS language, which is informal, slang-heavy, emoji-rich, and behaviorally grounded.

The original RedOne is a post-trained variant of **Qwen2.5**. It introduces no new architecture, tokenizer, or special SNS-specific modules; domain specificity comes from data composition and a three-stage training strategy: **Continue Pretraining (CPT)**, **Supervised Fine-Tuning (SFT)**, and **Preference Optimization (PO)**. The reported model sizes are **RedOne-7B**, built from **Qwen2.5-7B-Instruct**, and **RedOne-32B**, built from **Qwen2.5-32B** [2507.10605].

During **CPT**, the authors collect more than **100B tokens** from general high-quality corpora and SNS-specific domain data, then apply task-oriented rule filtering, **SLM filtering**, and **RegMix**, reducing the corpus to about **20B high-quality tokens**. Continue pretraining then uses standard causal language modeling with sequence length **4,096**, learning rate \(1 \times 10^{-5}\), for **1 epoch** [2507.10605].

The **SFT** stage constructs a large multi-task instruction suite from SNS notes, comments, queries, and interaction logs, while preserving SNS style such as informal wording, sarcasm, sentiment, and topic drift. Tasks are organized into six capability areas: **Content Understanding**, **Information Extraction**, **Semantic Matching**, **User Behavior Modeling**, **Dialogue and Persona Simulation**, and **Translation**. The paper explicitly names tasks such as **Note Taxonomy**, **Query Classification**, **Query Intent Recognition**, **Hashtag Prediction**, **Machine Reading Comprehension**, **Highlight Word Detection**, **Query-Note Relevance**, **Query-Note Retrieval**, **Post-View Search**, **Emotional Companion Dialogue**, **Role-playing Dialogue**, and **SNS Domain Translation**. General instruction data such as **LIMA**, **Infinity Instruct**, **WildChat**, **OpenThoughts**, and **LIMO** are mixed in to maintain general capabilities. SFT is performed in two steps: **3 epochs** on balanced SNS and general data, then **2 more epochs** with an SNS-heavier ratio, using sequence length up to **16,384**, batch size **128**, AdamW with \(\beta_1 = 0.9\), \(\beta_2 = 0.95\), \(\epsilon = 10^{-8}\), learning rate \(3 \times 10^{-6}\), and linear warmup ratio **0.1** [2507.10605].

The **PO** stage uses **Direct Preference Optimization (DPO)** on both subjective and objective SNS tasks. Preference pairs come from human expert judgments, judge-model annotations, explicit correct-versus-incorrect labels, and model-error constructions. Implementation details are **learning rate \(= 1 \times 10^{-7}\)**, **batch size \(= 64\)**, **sequence length \(= 4,096\)**, **epochs \(= 2\)**, and an auxiliary **SFT loss coefficient \(= 0.3\)** [2507.10605].

Evaluation covers **SNS-Bench**, containing **8 tasks**, and **SNS-TransBench**, a bilingual benchmark for **ZH→EN** and **EN→ZH** SNS translation. For **RedOne-7B**, the paper reports **General-Bench \(= 63.83\)**, **SNS-Bench \(= 66.88\)**, and **SNS-TransBench \(= 48.11\)**, compared with **Qwen2.5-7B** at **63.01 / 52.86 / 40.55**. The paper summarizes this as up to **+14.02% average improvement on SNS-Bench** and **+7.56%** on SNS-TransBench, with a slight **+0.82%** gain on general benchmarks [2507.10605]. For **RedOne-32B**, the reported numbers are **73.72 / 69.03 / 49.84**, versus **71.68 / 58.89 / 43.44** for **Qwen2.5-32B**, summarized as **+10.14%** on SNS-Bench, **+6.40%** on SNS-TransBench, and **+2.04%** on general benchmarks.

The paper also reports online A/B tests in two internal pipelines. In **Harmful Content Detection**, RedOne reduces **Exposure Rate** by **11.23%** relative to an existing single-task finetuned baseline. In **Post-View Search**, it improves **Click Page Rate** by **14.95%** [2507.10605]. Ablation results show that the full **CPT + SFT + PO** pipeline outperforms partial configurations, and comparisons against task-specific finetuning show that a domain-adapted RedOne provides a stronger base for further task specialization than the generic Qwen2.5 model.

## 6. RedOne 2.0: RL-prioritized post-training for SNS

**RedOne 2.0** is a redesigned successor to the original SNS RedOne, built by post-training the **Qwen3** family at **4B**, **8B**, **30B-A3B**, and **32B** scales with a **progressive, RL-prioritized three-stage pipeline** [2511.07070]. The redesign responds to a specific problem identified in the original RedOne recipe: an SFT-dominated pipeline can produce a “seesaw” between in-distribution SNS gains and out-of-distribution robustness, especially in smaller models.

The new pipeline consists of **Exploratory Learning**, **Targeted Fine-Tuning**, and **Refinement Learning**. The training corpus is normalized to \((Q, A)\) format and combines **SNS domain data** covering more than **75 tasks** with curated **general-domain data** containing rationales. Stage 1, **Exploratory Learning**, uses **DAPO-based RL** on approximately **750K** SNS entries and **50K** general entries. A defining feature is the use of **task-type-specific reward functions** rather than a single scalar signal. The paper specifies four types: **Exact Match** for closed-ended classification, **metrics-based rewards** such as BLEU or chrF++ for open-ended tasks like translation, **sandbox rewards** for executable tasks, and **pattern rewards** for format or instruction-following tasks [2511.07070].

The RL objective is a clipped-ratio DAPO loss over grouped candidates sampled from the old policy. The paper defines normalized token-level advantages from the sample rewards and optimizes a trust-region-style objective with clipping parameters \(\varepsilon_{\mathrm{low}} = 0.2\) and \(\varepsilon_{\mathrm{high}} = 0.28\). This stage is intended to establish initial SNS alignment while also diagnosing hard sub-tasks that remain weak after exploratory training [2511.07070].

Stage 2, **Targeted Fine-Tuning**, constructs a dataset of **1.8M examples**, comprising **1.7M SNS examples** focused on Stage-1 failure buckets and **100K general examples** with **soft labels**. For these general examples, the Stage-1 model generates **8 candidate outputs** per prompt, and a separate **judge model** selects the best candidate as the soft target. The stated purpose is to reduce distributional mismatch between SFT targets and the model’s own prior while regularizing against catastrophic forgetting. SFT uses standard next-token cross-entropy for **2 epochs**, batch size **64**, maximum sequence length **16,384**, optimizer **AdamW**, learning rate \(5 \times 10^{-6}\), warmup ratio **0.1**, and cosine decay [2511.07070].

Stage 3, **Refinement Learning**, applies DAPO RL again on about **400K examples** drawn from SNS and general sources, emphasizing difficult subsets and increasing the share of rationale-rich samples to **57.18%**. The RL stages use **500 steps**, maximum prompt length **10,000 tokens**, maximum response length **8,192 tokens**, an additional **4,096-token overlong buffer** with penalty factor **1.0**, prompt batch size **1,024**, **16 responses per prompt**, mini-batch size **256 prompts**, **4 gradient updates per rollout**, AdamW, constant learning rate \(5 \times 10^{-6}\), weight decay **0.1**, and linear warmup for **10 rollout steps** [2511.07070].

The quantitative comparisons emphasize compact-model efficiency. The abstract reports that the **4B RedOne 2.0** model delivers average improvements of about **2.41** over the **7B** sub-optimal baseline and average lift of about **8.74** from the **Qwen3-4B** base model, while using **less than half the domain data** required by the SFT-centric original RedOne [2511.07070]. In the stage ablation for **Qwen3-4B**, the base model scores **69.80 / 51.81 / 38.22** on **General-Bench / SNS-Bench / SNS-TransBench**; **Exploratory Learning only** yields **71.25 / 62.27 / 43.35**; **RL→SFT** yields **70.04 / 65.67 / 47.72**; and the full **RL→SFT→RL** pipeline reaches **70.80 / 67.57 / 47.67** [2511.07070]. The paper also reports that a naive **SFT→RL** baseline ends below the full RL-prioritized curriculum on all three aggregated benchmarks.

Task-level ablations reinforce the claim that one unified domain model can approach or exceed task-specific finetuning. For **Qwen3-4B**, **RedOne 2.0** scores **60.92** on **QCorr** versus **60.11** for task-specific fine-tuning, **66.54** on **MRC** versus **57.54**, **79.11** on **CHLW** versus **67.24**, and **47.67** on **SNS-TransBench** versus **44.25** [2511.07070]. The exceptions are also reported explicitly: for **Hashtag**, task-specific fine-tuning reaches **90.12** versus **89.05** for RedOne 2.0, and for **QGen**, task-specific fine-tuning reaches **49.24** versus **47.17**.

The paper also includes an online A/B experiment on **personalized re-creation of post titles** for a platform with **more than 3M users**. It reports **Advertiser Value (AdvV) ↑ 0.43%**, **Vague titles ratio ↓ 11.9%**, **Practical titles ratio ↑ 7.1%**, **Authentic titles ratio ↑ 12.9%**, and **Interactive titles ratio ↑ 25.8%** [2511.07070]. At the same time, a negative case study shows that style optimization can omit critical factual information, specifically a case where “transportation card” is lost, and the paper identifies this as a **faithfulness vs engagement** trade-off.

The limitations of RedOne 2.0 are correspondingly centered on reward design, safety, and operational cost. The paper notes dependence on curated task-specific rewards, limited treatment of bias and adversarial safety, incomplete coverage of evolving SNS task families, and the compute cost of RL with long context and large rollout batches [2511.07070]. A plausible implication is that RedOne 2.0 is best understood not as a finished universal SNS model but as a post-training template in which RL is placed at the center of domain adaptation rather than at the end of an SFT-heavy pipeline.

Source: https://www.emergentmind.com/topics/redone