ARXIV:2602.12222 · LLM TRAINING · SUBMITTED 02 APR · 02:30 UTC · FRESHNESS STALE

VerifiedSource: PDF linkedPartialPaperPack: 3 of 4 citation fields filledMissingMissing fields: authorsPartialProof: unverified proof status

Towards On-Policy SFT: Distribution Discriminant Theory and its Applications in LLM Training

arXiv

Develop a high-performance supervised fine-tuning framework based on Distribution Discriminant Theory to achieve RL-like generalization in computationally efficient settings.

Blocked on Code›Score5.0Evidence unverified

Opportunity summary

Pain Develop a high-performance supervised fine-tuning framework based on Distribution Discriminant Theory to achieve RL-like generalization in computationally efficient settings.

Evidence 0 refs | 0 sources | 17% coverage

Blocker Evidence unverified

Open Build Read PDF Signal Canvas Track

PROBLEM

Develop a high-performance supervised fine-tuning framework based on Distribution Discriminant Theory to achieve RL-like generalization in computationally efficient settings. This gap is primarily driven by RL's use of on-policy data.

METHOD

Supervised fine-tuning (SFT) is computationally efficient but often yields inferior generalization compared to reinforcement learning (RL). This gap is primarily driven by RL's use of on-policy data.

Full abstract

Supervised fine-tuning (SFT) is computationally efficient but often yields inferior generalization compared to reinforcement learning (RL). This gap is primarily driven by RL's use of on-policy data. We propose a framework to bridge this chasm by enabling On-Policy SFT. We first present \textbf{\textit{Distribution Discriminant Theory (DDT)}}, which explains and quantifies the alignment between data and the model-induced distribution. Leveraging DDT, we introduce two complementary techniques: (i) \textbf{\textit{In-Distribution Finetuning (IDFT)}}, a loss-level method to enhance generalization ability of SFT, and (ii) \textbf{\textit{Hinted Decoding}}, a data-level technique that can re-align the training corpus to the model's distribution. Extensive experiments demonstrate that our framework achieves generalization performance on par with prominent offline RL algorithms, including DPO and SimPO, while maintaining the efficiency of an SFT pipeline. The proposed framework thus offers a practical alternative in domains where RL is infeasible. We open-source the code here: https://github.com/zhangmiaosen2000/Towards-On-Policy-SFT

RESULT

ScienceToStartup currently rates this 5.0/10 on the public viability pass. Extensive experiments demonstrate that our framework achieves generalization performance on par with prominent offline RL algorithms, including DPO and SimPO, while maintaining the efficiency…

WHY NOW

LLM Training moved forward this cycle; last verified April 2026. Public score 5.0/10.

Continue into Read for claims, analysis, references, and neighboring papers.

Opportunity summary

Score5.0

PainDevelop a high-performance supervised fine-tuning framework based on Distribution Discriminant Theory to achieve RL-like generalization in computationally efficient settings.

Evidence0 refs | 0 sources | 17% coverage

Blockermissing authors

Analysis summary

Develop a high-performance supervised fine-tuning framework based on Distribution Discriminant Theory to achieve RL-like generalization in computationally efficient settings.

VerifiedSource: PDF linkedPartialPaperPack: 3 of 4 citation fields filledMissingMissing fields: authorsPartialProof: unverified proof status

ARXIV:2602.12222 · LLM TRAINING · SUBMITTED 02 APR · 02:30 UTC · FRESHNESS STALE

VerifiedSource: PDF linkedPartialPaperPack: 3 of 4 citation fields filledMissingMissing fields: authorsPartialProof: unverified proof status

Towards On-Policy SFT: Distribution Discriminant Theory and its Applications in LLM Training

arXiv

Develop a high-performance supervised fine-tuning framework based on Distribution Discriminant Theory to achieve RL-like generalization in computationally efficient settings.

Blocked on Code›Score5.0Evidence unverified

Opportunity summary

Pain Develop a high-performance supervised fine-tuning framework based on Distribution Discriminant Theory to achieve RL-like generalization in computationally efficient settings.

Evidence 0 refs | 0 sources | 17% coverage

Blocker Evidence unverified

Open Build Read PDF Signal Canvas Track

PROBLEM

METHOD

Supervised fine-tuning (SFT) is computationally efficient but often yields inferior generalization compared to reinforcement learning (RL). This gap is primarily driven by RL's use of on-policy data.

Full abstract

RESULT

WHY NOW

LLM Training moved forward this cycle; last verified April 2026. Public score 5.0/10.

Continue into Read for claims, analysis, references, and neighboring papers.

Opportunity summary

Score5.0

PainDevelop a high-performance supervised fine-tuning framework based on Distribution Discriminant Theory to achieve RL-like generalization in computationally efficient settings.

Evidence0 refs | 0 sources | 17% coverage

Blockermissing authors

Analysis summary

Develop a high-performance supervised fine-tuning framework based on Distribution Discriminant Theory to achieve RL-like generalization in computationally efficient settings.

VerifiedSource: PDF linkedPartialPaperPack: 3 of 4 citation fields filledMissingMissing fields: authorsPartialProof: unverified proof status

Paper Pack

10.48550/arXiv.2602.12222

Towards On-Policy SFT: Distribution Discriminant Theory and its Applications in LLM Training

Develop a high-performance supervised fine-tuning framework based on Distribution Discriminant Theory to achieve RL-like generalization in computationally efficient settings.

Abstract

Source availability

PDF linked

The paper record includes a public PDF URL.

Extraction status

Derived fallback

Read summaries are estimated from adjacent metadata, not verified extraction rows.

Proof status

unverified

0 refs; 0 sources; 17% coverage.

What was readable

linkedon filenot materializedderived fallback37 indexednot indexed

Derived fallback: Estimated from adjacent evidence; not verified from source.

Viability

5.0

Time to MVP

MVP estimate missing

Commercial

No commercial flags on file

Export

Preparing verified analysis

lens / founder

PROBLEM

METHOD

Supervised fine-tuning (SFT) is computationally efficient but often yields inferior generalization compared to reinforcement learning (RL). This gap is primarily driven by RL's use of on-policy data.

RESULT

WHY NOW

LLM Training moved forward this cycle; last verified April 2026. Public score 5.0/10.

Claim map

Abstract-backed public claims while anchored extraction refreshes.

Strong 0Mixed 0Weak 4

Evidencepartial
Develop a high-performance supervised fine-tuning framework based on Distribution Discriminant Theory to achieve RL-like generalization in computationally efficient settings. This gap is primarily driven by RL's use of on-policy data.
Implicationpartial
Abstract-backed fallback claim; anchored extraction has not materialized a public claim row yet.
Verificationpartial
partial
Evidencepartial
Supervised fine-tuning (SFT) is computationally efficient but often yields inferior generalization compared to reinforcement learning (RL). This gap is primarily driven by RL's use of on-policy data.
Implicationpartial
Abstract-backed fallback claim; anchored extraction has not materialized a public claim row yet.
Verificationpartial
partial
Evidencepartial
ScienceToStartup currently rates this 5.0/10 on the public viability pass. Extensive experiments demonstrate that our framework achieves generalization performance on par with prominent offline RL algorithms, including DPO and SimPO, while maintaining the efficiency of an SFT pipeline.
Implicationpartial
Abstract-backed fallback claim; anchored extraction has not materialized a public claim row yet.
Verificationpartial
partial
Evidencepartial
LLM Training moved forward this cycle; last verified April 2026. Public score 5.0/10.
Implicationpartial
Abstract-backed fallback claim; anchored extraction has not materialized a public claim row yet.
Verificationpartial
partial

Constellation map

Paper-native neighborhood for concepts, methods, materials, markets, and competitors. Missing lanes stay labeled instead of disappearing behind commercialization gates.

Open full Signal Canvas

Concepts

not indexed

Methods

Supervised fine-tuning (SFT) is computationally efficient but often yields inferior generalization compared to reinforcement learning (RL). This gap is primarily driven by RL's use of on-policy data.

Materials

PDF linked

Markets

LLM Training

Competitors

not indexed

Competitive landscape

Develop a high-performance supervised fine-tuning framework based on Distribution Discriminant Theory to achieve RL-like generalization in computationally efficient settings.

Segment

LLM Training

Adoption evidence

No public code link in the paper record yet

Commercial read

5.0/10 public viability

Direct

not classified

Adjacent

not classified

Substitute

not classified

Unknown

not classified

Buzz

No indexed public discussion is attached to 2602.12222 yet. That is a visibility signal, not a blank module: the monitor is watching the public channels below.

Hacker News

Not indexed yet

Bluesky

Not indexed yet

PDF

Preview the source document here, or use the hero PDF action for a new tab.

References(37)

Reinforcement Learning via Self-Distillation

2026Jonas Hubotter, Frederike Lubeck et al.

Self-Distillation Enables Continual Learning

2026Idan Shenfeld, Mehul Damani et al.

Self-Distilled Reasoner: On-Policy Self-Distillation for Large Language Models

2026Siyan Zhao, Zhihui Xie et al.

ProFit: Leveraging High-Value Signals in SFT via Probability-Guided Token Selection

2026Tao Liu, Taiqiang Wu et al.

Entropy-Adaptive Fine-Tuning: Resolving Confident Conflicts to Mitigate Forgetting

2026Muxi Diao, Lele Yang et al.

Mitigating Forgetting Between Supervised and Reinforcement Learning Yields Stronger Reasoners

2025Xiangchi Yuan, Xiang Chen et al.

Debunk the Myth of SFT Generalization

2025Xiaofeng Lin, Hejian Sang et al.

Anchored Supervised Fine-Tuning

2025He Zhu, Junyou Su et al.

On the Generalization of SFT: A Reinforcement Learning Perspective with Reward Rectification

2025Yongliang Wu, Yizhou Zhou et al.

Qwen3 Technical Report

2025An Yang, Anfeng Li et al.

Large Language Diffusion Models

2025Shen Nie, Fengqi Zhu et al.

SFT Memorizes, RL Generalizes: A Comparative Study of Foundation Model Post-training

2025Tianzhe Chu, Yuexiang Zhai et al.

The Llama 3 Herd of Models

2024Abhimanyu Dubey, Abhinav Jauhri et al.

SimPO: Simple Preference Optimization with a Reference-Free Reward

2024Yu Meng, Mengzhou Xia et al.

MathScale: Scaling Instruction Tuning for Mathematical Reasoning

2024Zhengyang Tang, Xingxing Zhang et al.

OlympiadBench: A Challenging Benchmark for Promoting AGI with Olympiad-Level Bilingual Multimodal Scientific Problems

2024Chaoqun He, Renjie Luo et al.

Relative Preference Optimization: Enhancing LLM Alignment through Contrasting Responses across Identical and Diverse Prompts

2024Yueqin Yin, Zhendong Wang et al.

Contrastive Preference Optimization: Pushing the Boundaries of LLM Performance in Machine Translation

2024Haoran Xu, Amr Sharaf et al.

Understanding the Effects of RLHF on LLM Generalisation and Diversity

2023Robert Kirk, Ishita Mediratta et al.

Cumulative Reasoning with Large Language Models

2023Yifan Zhang, Jingqin Yang et al.

Showing 20 of 37 references

CITED BY

No citing papers are indexed in the public S2S graph yet. This is an explicit zero-signal state, not a hidden lookup.

Foundation

Prior WorkDeReason: A Difficulty-Aware Curriculum Improves Decoupled SFT-then-RL Training for General Reasoning

5.0

Extension

none indexed

Commercially relevant

Higher ViabilityDEFT: Distribution-guided Efficient Fine-Tuning for Human Alignment

7.0

Higher ViabilityEntropy-KL Divergence-based Token Masking: A Novel Approach for Selective Fine-tuning of Large Language Models

7.0

Higher ViabilityVLA-OPD: Bridging Offline SFT and Online RL for Vision-Language-Action Models via On-Policy Distillation

7.0

Higher ViabilityPointRFT: Explicit Reinforcement Fine-tuning for Point Cloud Few-shot Learning

7.0

Higher ViabilityRobust Post-Training for Generative Recommenders: Why Exponential Reward-Weighted SFT Outperforms RLHF

7.0

Higher ViabilityUnlocking Data Value in Finance: A Study on Distillation and Difficulty-Aware Training

8.0

Higher ViabilitymSFT: Addressing Dataset Mixtures Overfiting Heterogeneously in Multi-task SFT

7.0

Conflicting

Competing ApproachComplementing reinforcement learning with SFT through logit averaging in the post training of LLMs

3.0

Competing ApproachLarge Language Model Post-Training: A Unified View of Off-Policy and On-Policy Learning

1.0

Related Resources

Owned Distribution

Subscribe to the weekly brief

Get the weekly shortlist of commercializable papers, benchmark movers, and proof receipts that matter for product execution.

Agent drawer

5 surfaces preserved for agents. Humans can ignore.

Developer contracts, payload previews, evidence maps, and run controls stay here instead of the Read, Build, and Track workspace.

Run context

Paper: 2602.12222
Route: /paper/towards-on-policy-sft-distribution-discriminant-theory-and-its-applications-in-llm-training
Active tab: read
Artifact: towards-on-policy-sft-distribution-discriminant-theory-and-its-applications-in-llm-training

Available agents

Read extractor
Build planner
Track monitor
Competitive mapper
Related-paper scout

API/MCP endpoints

REST paper pack API/api/v1/paper/towards-on-policy-sft-distribution-discriminant-theory-and-its-applications-in-llm-training/paper-pack
REST build passport API/api/v1/paper/towards-on-policy-sft-distribution-discriminant-theory-and-its-applications-in-llm-training/build-passport
REST OpenAPI/api/openapi.json
MCP descriptor/api/mcp
MCP resourcesciencetostartup://surfaces/paper-workspace

Tool contracts

paper_packbuild_passportopportunity_kernelforesightsource_proofevidence_state

Payload preview

Inspect payload

{
  "contract_version": "paper-r2",
  "paper_id": "6a39aa2f-ccce-42bf-8abd-5c7c22cd89de",
  "arxiv_id": "2602.12222",
  "canonical_route": "/paper/towards-on-policy-sft-distribution-discriminant-theory-and-its-applications-in-llm-training",
  "active_tab": "synced from current hash by the drawer client",
  "selected_artifact": "towards-on-policy-sft-distribution-discriminant-theory-and-its-applications-in-llm-training",
  "endpoints": {
    "paper_pack": "/api/v1/paper/towards-on-policy-sft-distribution-discriminant-theory-and-its-applications-in-llm-training/paper-pack",
    "build_passport": "/api/v1/paper/towards-on-policy-sft-distribution-discriminant-theory-and-its-applications-in-llm-training/build-passport",
    "mcp_resource": "sciencetostartup://surfaces/paper-workspace"
  }
}

Schema validation

paper-r2 contract: present
JSON-LD twin: SSR emitted
OpenAPI path parity: /api/openapi.json
MCP resource parity: paper-workspace

Job trace

queued: drawer opened by user action
running: inspect or copy payload
succeeded: payload available in SSR
failed: route errors appear in evidence cards

Evidence map

sources used: page freshness, source proof anchors, JSON-LD
missing sources: exposed by PaperPack and EvidenceState chips
derived fallbacks: marked unverified before handoff

Page Freshness

Canonical route, proof status, last verified, refs, sources, and coverage.

Page Freshness

Paper proof surface

Canonical route: /paper/towards-on-policy-sft-distribution-discriminant-theory-and-its-applications-in-llm-training

stale

Proof freshness: stale
Proof status: unverified
Display score: 5/10
Last proof check: 2026-04-02
Score updated: 2026-04-02
Score fresh until: 2026-05-02
References: 0
Source count: 0
Coverage: 17%

This page is showing the last landed evidence receipt and score bundle because the latest proof data is outside the freshness window.

OpenAlex: pending — this preprint is not yet indexed by OpenAlex.

Agent Handoff

Endpoint list, payload shape, route context, and copyable handoff data.

Agent Handoff

Towards On-Policy SFT: Distribution Discriminant Theory and its Applications in LLM Training

Canonical ID towards-on-policy-sft-distribution-discriminant-theory-and-its-applications-in-llm-training | Route /paper/towards-on-policy-sft-distribution-discriminant-theory-and-its-applications-in-llm-training

REST example

curl https://sciencetostartup.com/api/v1/agent-handoff/paper/towards-on-policy-sft-distribution-discriminant-theory-and-its-applications-in-llm-training

MCP example

{
  "tool": "get_paper",
  "arguments": {
    "arxiv_id": "2602.12222"
  }
}

source_context

{
  "surface": "paper",
  "mode": "paper",
  "query": "Towards On-Policy SFT: Distribution Discriminant Theory and its Applications in LLM Training",
  "normalized_query": "2602.12222",
  "route": "/paper/towards-on-policy-sft-distribution-discriminant-theory-and-its-applications-in-llm-training",
  "paper_ref": "towards-on-policy-sft-distribution-discriminant-theory-and-its-applications-in-llm-training",
  "topic_slug": null,
  "benchmark_ref": null,
  "dataset_ref": null
}

Buildability Receipt

Verdict, compute envelope, blockers, signature state, and receipt links.

Paper proof page receipt window

Watch and verify: Towards On-Policy SFT: Distribution Discriminant Theory and its Applications in LLM Training

/buildability/towards-on-policy-sft-distribution-discriminant-theory-and-its-applications-in-llm-training

Watchwatch

Subject: Towards On-Policy SFT: Distribution Discriminant Theory and its Applications in LLM Training

Verdict

Watch

Verdict is Watch because viability or proof quality is intermediate and should be re-evaluated before execution.

Time to first demo

Insufficient data

No first-demo timestamp, owner estimate, or elapsed demo receipt is attached to this surface.

Compute envelope

Structured compute envelope

Insufficient data

No data, compute, hardware, memory, latency, dependency, or serving requirement receipt is attached.

Evidence ids

Receipt path

/buildability/towards-on-policy-sft-distribution-discriminant-theory-and-its-applications-in-llm-training

Paper ref

towards-on-policy-sft-distribution-discriminant-theory-and-its-applications-in-llm-training

arXiv id

2602.12222

Freshness

Generated at

2026-04-02T02:30:40.136Z

Evidence freshness

stale

Last verification

2026-04-02T02:30:40.136Z

Sources

References

Coverage

17%

Hash state

Lineage hash

f4697cbb5d8f26e0bc9841cd97671becb817d4cab5ef25c583e5274393d4a1b5

Canonical opportunity-kernel lineage hash.

Signature state

External signature

unsigned_external

No founder, registry, pilot, or production-adoption signature is attached to this receipt.

Verification

not_verified

Verification is blocked until an external signature is provided.

Blockers

Missing: repo_url
Missing: references
Missing: proof_status
Missing: distribution_readiness_scores
Missing: paper_extraction_scorecards
Unknown: distribution readiness has not been computed yet
Unknown: proof verification has not been recorded yet

Verification pending / evidence receipt incomplete

repo_url

references

Missing proof, requirement, signature, approval, adoption, or telemetry fields are blockers and must not be inferred.

Open receipt API receipt Build Loop Signal Canvas Proof divergence Divergence API Brier outcomes API

Source Proof anchors

Visual citations from the paper document graph.

JSON-LD twin

The application/ld+json payload rendered for agents.

{
  "@context": "https://schema.org",
  "@graph": [
    {
      "@type": "WebPage",
      "@id": "https://sciencetostartup.com/paper/towards-on-policy-sft-distribution-discriminant-theory-and-its-applications-in-llm-training#webpage",
      "url": "https://sciencetostartup.com/paper/towards-on-policy-sft-distribution-discriminant-theory-and-its-applications-in-llm-training",
      "name": "Towards On-Policy SFT: Distribution Discriminant Theory and its Applications in LLM Training",
      "description": "Develop a high-performance supervised fine-tuning framework based on Distribution Discriminant Theory to achieve RL-like generalization in computationally efficient settings.",
      "isPartOf": {
        "@id": "https://sciencetostartup.com/#website"
      }
    },
    {
      "@type": "ScholarlyArticle",
      "@id": "https://sciencetostartup.com/paper/towards-on-policy-sft-distribution-discriminant-theory-and-its-applications-in-llm-training#scholarlyArticle",
      "headline": "Towards On-Policy SFT: Distribution Discriminant Theory and its Applications in LLM Training",
      "description": "Develop a high-performance supervised fine-tuning framework based on Distribution Discriminant Theory to achieve RL-like generalization in computationally efficient settings.",
      "url": "https://sciencetostartup.com/paper/towards-on-policy-sft-distribution-discriminant-theory-and-its-applications-in-llm-training",
      "sameAs": "https://arxiv.org/abs/2602.12222",
      "identifier": {
        "@type": "PropertyValue",
        "propertyID": "arXiv",
        "value": "2602.12222"
      },
      "isAccessibleForFree": true,
      "isPartOf": {
        "@id": "https://sciencetostartup.com/#website"
      },
      "datePublished": "2026-02-12T17:59:58.000Z",
      "citation": [
        {
          "@type": "ScholarlyArticle",
          "identifier": {
            "@type": "PropertyValue",
            "propertyID": "SemanticScholar",
            "value": "89e83ee3eb95843292b1f75f0e50b89e7667b828"
          },
          "url": "https://www.semanticscholar.org/paper/89e83ee3eb95843292b1f75f0e50b89e7667b828"
        },
        {
          "@type": "ScholarlyArticle",
          "identifier": {
            "@type": "PropertyValue",
            "propertyID": "SemanticScholar",
            "value": "370a293b5dcf9d456ba7f805b5b0a3d8e984ff6e"
          },
          "url": "https://www.semanticscholar.org/paper/370a293b5dcf9d456ba7f805b5b0a3d8e984ff6e"
        },
        {
          "@type": "ScholarlyArticle",
          "identifier": {
            "@type": "PropertyValue",
            "propertyID": "SemanticScholar",
            "value": "06571d39f14fbd771423bf57c9a237faaf7994bf"
          },
          "url": "https://www.semanticscholar.org/paper/06571d39f14fbd771423bf57c9a237faaf7994bf"
        },
        {
          "@type": "ScholarlyArticle",
          "identifier": {
            "@type": "PropertyValue",
            "propertyID": "SemanticScholar",
            "value": "215c0c04dcd378aedbc59187956dac2ae5bc446c"
          },
          "url": "https://www.semanticscholar.org/paper/215c0c04dcd378aedbc59187956dac2ae5bc446c"
        },
        {
          "@type": "ScholarlyArticle",
          "identifier": {
            "@type": "PropertyValue",
            "propertyID": "SemanticScholar",
            "value": "9bd7bd79c9daa98c31080bec43582f0498f81186"
          },
          "url": "https://www.semanticscholar.org/paper/9bd7bd79c9daa98c31080bec43582f0498f81186"
        },
        {
          "@type": "ScholarlyArticle",
          "identifier": {
            "@type": "PropertyValue",
            "propertyID": "SemanticScholar",
            "value": "4ba74db3b4c287fd4a432e520c79118bbd16d28b"
          },
          "url": "https://www.semanticscholar.org/paper/4ba74db3b4c287fd4a432e520c79118bbd16d28b"
        },
        {
          "@type": "ScholarlyArticle",
          "identifier": {
            "@type": "PropertyValue",
            "propertyID": "SemanticScholar",
            "value": "ab6265c7047c3e1ea0125a23303d83d8b37c8809"
          },
          "url": "https://www.semanticscholar.org/paper/ab6265c7047c3e1ea0125a23303d83d8b37c8809"
        },
        {
          "@type": "ScholarlyArticle",
          "identifier": {
            "@type": "PropertyValue",
            "propertyID": "SemanticScholar",
            "value": "81696e83b871ad9c24109cba8f79d1b75003f0f0"
          },
          "url": "https://www.semanticscholar.org/paper/81696e83b871ad9c24109cba8f79d1b75003f0f0"
        },
        {
          "@type": "ScholarlyArticle",
          "identifier": {
            "@type": "PropertyValue",
            "propertyID": "SemanticScholar",
            "value": "1414653547b72a23d2c0ca987d270fb236c9105a"
          },
          "url": "https://www.semanticscholar.org/paper/1414653547b72a23d2c0ca987d270fb236c9105a"
        },
        {
          "@type": "ScholarlyArticle",
          "identifier": {
            "@type": "PropertyValue",
            "propertyID": "SemanticScholar",
            "value": "d2d84d56f730f81d276a02b48d5d44db5bde0b4a"
          },
          "url": "https://www.semanticscholar.org/paper/d2d84d56f730f81d276a02b48d5d44db5bde0b4a"
        },
        {
          "@type": "ScholarlyArticle",
          "identifier": {
            "@type": "PropertyValue",
            "propertyID": "SemanticScholar",
            "value": "0d11a9674b68216b92e08cf7617a93fbd3fb91f4"
          },
          "url": "https://www.semanticscholar.org/paper/0d11a9674b68216b92e08cf7617a93fbd3fb91f4"
        },
        {
          "@type": "ScholarlyArticle",
          "identifier": {
            "@type": "PropertyValue",
            "propertyID": "SemanticScholar",
            "value": "a99dee9602e21a71526b9681d8dba37c55b66941"
          },
          "url": "https://www.semanticscholar.org/paper/a99dee9602e21a71526b9681d8dba37c55b66941"
        },
        {
          "@type": "ScholarlyArticle",
          "identifier": {
            "@type": "PropertyValue",
            "propertyID": "SemanticScholar",
            "value": "40e8af970329135ec95057d73e239dab805ad128"
          },
          "url": "https://www.semanticscholar.org/paper/40e8af970329135ec95057d73e239dab805ad128"
        },
        {
          "@type": "ScholarlyArticle",
          "identifier": {
            "@type": "PropertyValue",
            "propertyID": "SemanticScholar",
            "value": "c3f1fae241a3c2449e675ab750873d800f95513c"
          },
          "url": "https://www.semanticscholar.org/paper/c3f1fae241a3c2449e675ab750873d800f95513c"
        },
        {
          "@type": "ScholarlyArticle",
          "identifier": {
            "@type": "PropertyValue",
            "propertyID": "SemanticScholar",
            "value": "3352782f94354d3f3a170f497dd1888e9cd39d8a"
          },
          "url": "https://www.semanticscholar.org/paper/3352782f94354d3f3a170f497dd1888e9cd39d8a"
        },
        {
          "@type": "ScholarlyArticle",
          "identifier": {
            "@type": "PropertyValue",
            "propertyID": "SemanticScholar",
            "value": "bcf2c7e3f4ed64c8294c35a59220a26dd4f40060"
          },
          "url": "https://www.semanticscholar.org/paper/bcf2c7e3f4ed64c8294c35a59220a26dd4f40060"
        },
        {
          "@type": "ScholarlyArticle",
          "identifier": {
            "@type": "PropertyValue",
            "propertyID": "SemanticScholar",
            "value": "28d5e9fb1190e0cd0efcc8206bb73d7112904910"
          },
          "url": "https://www.semanticscholar.org/paper/28d5e9fb1190e0cd0efcc8206bb73d7112904910"
        },
        {
          "@type": "ScholarlyArticle",
          "identifier": {
            "@type": "PropertyValue",
            "propertyID": "SemanticScholar",
            "value": "ebd1c04c61f73f46def3305ca11d038c46665b65"
          },
          "url": "https://www.semanticscholar.org/paper/ebd1c04c61f73f46def3305ca11d038c46665b65"
        },
        {
          "@type": "ScholarlyArticle",
          "identifier": {
            "@type": "PropertyValue",
            "propertyID": "SemanticScholar",
            "value": "cb3968152f7d93f53d24b00279a90d5071ddc85a"
          },
          "url": "https://www.semanticscholar.org/paper/cb3968152f7d93f53d24b00279a90d5071ddc85a"
        },
        {
          "@type": "ScholarlyArticle",
          "identifier": {
            "@type": "PropertyValue",
            "propertyID": "SemanticScholar",
            "value": "507acddb0b7f36b83fd7c8bff2f121eb506ac8fb"
          },
          "url": "https://www.semanticscholar.org/paper/507acddb0b7f36b83fd7c8bff2f121eb506ac8fb"
        }
      ],
      "additionalProperty": [
        {
          "@type": "PropertyValue",
          "propertyID": "viabilityScore",
          "value": 5
        },
        {
          "@type": "PropertyValue",
          "propertyID": "researchDomain",
          "value": "LLM Training"
        }
      ]
    },
    {
      "@type": "BreadcrumbList",
      "itemListElement": [
        {
          "@type": "ListItem",
          "position": 1,
          "name": "Home",
          "item": "https://sciencetostartup.com"
        },
        {
          "@type": "ListItem",
          "position": 2,
          "name": "LLM Training",
          "item": "https://sciencetostartup.com/topics"
        },
        {
          "@type": "ListItem",
          "position": 3,
          "name": "Towards On-Policy SFT: Distribution Discriminant Theory and ",
          "item": "https://sciencetostartup.com/paper/towards-on-policy-sft-distribution-discriminant-theory-and-its-applications-in-llm-training"
        }
      ]
    }
  ]
}

Towards On-Policy SFT: Distribution Discriminant Theory and its Applications in LLM Training

Towards On-Policy SFT: Distribution Discriminant Theory and its Applications in LLM Training

Claim map

Constellation map

Competitive landscape

Buzz

PDF

References(37)

Related Papers

Related Resources

Subscribe to the weekly brief

Build artifacts

Brief

Experiment plan

Validation checklist

Scientific founder

Translational engineer

Domain operator

GTM lead

Regulatory/clinical advisor

Timeline

Claim map

Constellation map

Competitive landscape

Buzz

PDF

References(37)

Related Papers

Related Resources

Subscribe to the weekly brief

Build artifacts

Brief

Experiment plan

Validation checklist

Scientific founder

Translational engineer

Domain operator

GTM lead

Regulatory/clinical advisor

Timeline