DiRAFT is a method for generating privacy‑aware financial transcripts by extracting concept graphs from customer conversations and sampling new graph plans from aggregate transition, starting‑concept, and plan‑length priors. The approach uses a discourse layer to realize text without referencing individual source passages, aiming to preserve discourse fidelity while reducing residual confidentiality risks. The authors evaluate privacy through corpus‑level structural overlap and source‑grounded checks for non‑public investor facts, finding that stochastic sampling yields lower structural overlap and fewer confidential fact matches compared to greedy or top‑3 strategies.
arXiv:2602. 07294v4 Announce Type: replace-cross Abstract: With the increasing deployment of Large Language Models (LLMs) in the finance domain, LLMs are increasingly expected to parse complex regulatory disclosures.
arXiv:2606. 18192v1 Announce Type: new Abstract: As high-quality public web corpora become increasingly exhausted, clean long-context documents have become a scarce and expensive source of training data for large language models (LLMs).
arXiv:2610.19240v1 Announce Type: new Abstract: Financial disclosures express the underlying relationships of double-entry accounting through unstructured text. While financial LLM benchmarks predomi...
The paper introduces a benchmark for evaluating large language models on trustworthy analysis of earnings call transcripts. It proposes a numeric evidence evaluation method that assesses groundedness without expert annotation, and presents an automated pipeline that builds the ECTs-100 dataset from the top 100 S&P 500 constituents. The study also explores the failure mode of conscious incompetence, where models must recognize insufficient evidence and avoid hallucinations, finding that while groundedness is strong, correctness remains a challenge.
arXiv:2607. 14174v1 Announce Type: new Abstract: Financial sentiment extraction has largely relied on news text and supervised extraction against return labels alone, leaving 10-K filings -- and volatility, the target risk disclosure is arguably best suited to informing -- comparatively unexplored.
The paper introduces a training‑free, alignment‑free method for corporate intelligence that uses deterministic sparse seed vectors to hash word strings into a fixed high‑dimensional basis. By accumulating these seed vectors across sentence contexts, the authors create corpus‑specific semantic signatures that enable rapid document comparison, issuer fingerprinting, vocabulary shift tracking, and thematic sentence extraction—all on standard CPU hardware. Applied to a multi‑year set of SEC filings, the approach reveals distinct semantic profiles for major corporate events such as Boeing’s 737 MAX crisis, Intel’s supply‑chain disruptions, and Bunge’s acquisition of Viterra, with each profile traceable to its source sentences without any domain‑specific training or LLM inference.
FinMultiHopQA is a new benchmark comprising 15,600 financial questions that require multi‑hop reasoning across filing sections, fiscal years, and peer companies. It is grounded in FinReflectKG, a temporally indexed financial knowledge graph with over 17.5 million triples derived from SEC 10‑K filings. The benchmark uniquely supports controlled multi‑hop taxonomies, document topologies, motif‑level reproducibility, and provenance‑linked auditability, and it allows researchers to test how different evidence presentations affect large language model performance.
FIRME is a registry‑anchored memory graph designed to integrate investor knowledge across multiple sources and time periods while preserving identity, provenance, and context. It reconciles varying surface forms of observations into canonical relationships, fuses investor‑call discussions with portfolio cash‑flow signals within periods, and uses persistent aliases to maintain consistent relationship histories across periods. Evaluation on proprietary investor‑call data and S&P 500 earnings‑call transcripts shows significant improvements in Correct Match Rate and reattachment‑recall compared to a baseline registry.
Financial disclosures contain numerical claims, temporal statements, entity references, policy commitments, and risk descriptions that may conflict in qualitatively different ways. Detecting a conflict is only the first step: review workflows may also need to determine its type, since numerical, temporal, referential, factual, and normative inconsistencies require different evidence and downstream checks.
arXiv:2608.24842v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly deployed as AI analysts to process financial disclosures and support AI-assisted investment decisions....
While corporate narrative disclosures provide crucial information to capital markets, comprehensively evaluating their qualitative changes over time remains challenging. Narrative text is inherently multidimensional, meaning that an improvement in one textual dimension often occurs alongside changes in others.
FIRME (Fusion and Identity Registry for Memory Evolution) is a registry‑anchored memory graph designed to integrate investor observations across time and sources while preserving identity, provenance, and context. It reconciles varying surface forms of recurring observations into canonical relationships, fuses investor‑call discussions with portfolio cash‑flow signals within periods, and uses persistent aliases to maintain consistent relationship histories across periods. Evaluations on proprietary investor‑call data, cash‑flow records, and S&P 500 earnings‑call transcripts show that FIRME improves Correct Match Rate by 7.8–16.5 percentage points and significantly enhances reattachment recall and merge precision compared to a baseline registry lacking alias memory.