Public Company Risks

Predict which 10-K risk factors actually precede enforcement, restatements, and drawdowns.

What we did

  • Scraped SEC EDGAR for 10-K, 10-Q, and 8-K filings with precise timestamps.
  • Generated forward-looking risk questions anchored at each filing date.
  • Labeled with actual downstream outcomes — enforcement actions, restatements, drawdowns.
  • Fine-tuned Qwen3-32B and benchmarked on 6,109 held-out SEC risk queries.

Example datapoint

A sample training example — question, source, and outcome-derived label.

DATASET SEC risk signals

RL

Question
Will Northwind Logistics Inc. (NASDAQ: NWIN) face a Commission enforcement action within 12 months of the Feb 28, 2025 10-K disclosure of an informal revenue-recognition inquiry?
Question source
10-K risk factors Feb 28, 2025
Disclosure of ongoing informal inquiry related to revenue recognition policies
Label
Yes.
Type
binary
Confidence
0.93
Label source
SEC litigation release Nov 4, 2025
Commission orders cease-and-desist and civil penalty in accounting fraud matter

DATASET SEC risk signals

RL
Question
Does Horizon Devices’ Mar 21, 2025 10-K (filed 4:32pm ET) contain Item 9A language on material ICFR weakness that predicts a restatement of FY2024 revenue within the next four quarters?
Question source
10-K MD&A Mar 21, 2025
Management cites material weakness in ICFR; auditor expanded scope
Label
No.
Type
binary
Confidence
0.81
Label source
8-K Aug 1, 2025
Annual report filed without revision to prior periods

DATASET SEC risk signals

RL
Question
What is the probability that S&P Global Ratings lowers Meridian Rail Corp.’s senior unsecured rating from BBB to BBB- or below before Jun 30, 2025, given May 2 CDS at 285 bp and net leverage 4.4× per the 8-K?
Question source
8-K May 2, 2025
Company completes $900m covenant-lite add-on; leverage pro forma 4.4×
Label
0.41
Type
continuous
Confidence
0.84
Label source
S&P Global Jun 28, 2025
Affirm BBB/A-2 outlook revised to negative; no notch change in quarter

DATASET SEC risk signals

RL
Question
Among Item 1A risk themes in S&P 500 industrial 10-Ks filed Q1 2025, which factor—(A) litigation, (B) cyber, (C) supply chain—best predicts a subsequent >5% peak-to-trough equity drawdown within 90 trading days?
Question source
SEC EDGAR Mar 31, 2025
Batch scrape: 58 industrial filers with standardized 1A headings
Label
Supply chain.
Type
binary
Confidence
0.72
Label source
Internal factor study Jul 15, 2025
Marginal lift for supply-chain-heavy 1A vs. peers on drawdown AUC

Results

Benchmark comparisons against frontier models.

SEC Risk Prediction: Brier Score, Skill, and Calibration

Fine-tuned Qwen3-32B achieves Brier Skill Score +11.6% with ECE of 0.029 across 6,109 SEC risk queries — 64.7% lower calibration error than GPT-5 (ECE 0.081). The model learns to distinguish boilerplate legal language from meaningful signals preceding adverse outcomes.