Pith. sign in

REVIEW 3 major objections 6 minor 21 references

From Financial Sentiment Classification to Return Predictability: A QLoRA Benchmark of Large Language Models

T0 review · 3 major / 6 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read QLoRA fine-tuning sharply improves financial sentiment classification, but the tuned models' daily sentiment scores do not reliably predict stock returns after statistical correction.

desk verdict A careful, honestly hedged benchmark paper: the QLoRA classification results are solid and useful, while the economic null is credible under its next-session design but weaker than the abstract's 'clear gap' phrasing suggests. read the letter →

arxiv 2608.04200 v1 pith:SRHIM4GM submitted 2026-08-04 q-fin.MF cs.LG

Fusheng Luo This is my paper · ORCID
classification q-fin.MFcs.LG
keywords financialsentimentanalysisQLoRAparameter-efficientfine-tuninglargelanguagemodelsreturnpredictabilityrankinformationcoefficientcross-sectionalstockreturnsnews
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper separates two questions that are often conflated: whether QLoRA fine-tuning makes 7-8B language models accurate financial sentiment classifiers, and whether that accuracy becomes cross-sectional return predictability. On a unified three-class benchmark combining five financial text sources, QLoRA raises Qwen2.5's macro-F1 from 0.7274 to 0.8615, and the best tuned backbone (Mistral-7B) reaches 0.8840 accuracy and 0.8771 macro-F1. On a temporally separate 2019 sample of 10,637 headlines linked to S&P 100 stocks, all seven probability-producing classifiers show positive but small one-day rank information coefficients, the largest being FinBERT's 0.0143, and none of the 28 model-horizon tests survives autocorrelation-consistent inference with false-discovery-rate correction. The paper concludes that parameter-efficient adaptation is effective for label accuracy while documenting a clear gap between semantic accuracy and tradable market signals.

What carries the argument

The carrying mechanism is QLoRA (a frozen 4-bit quantized backbone trained only through low-rank adapters), converted into a sequence-classification head from which each headline yields a continuous expected-polarity score $s = p_{\mathrm{pos}} - p_{\mathrm{neg}} \in [-1,1]$. Scores are averaged by stock and calendar date for stocks with fresh news, aligned to the first trading session strictly after the signal date with return $R^{(h)}_{i,t} = P^{\mathrm{close}}_{i,t+h-1}/P^{\mathrm{open}}_{i,t} - 1$, and then ranked cross-sectionally each day. Rank IC means are tested with heteroskedasticity- and autocorrelation-consistent standard errors and false-discovery-rate correction, and overlapping cohorts convert multi-day holdings into a daily portfolio return series. This pipeline is what turns label accuracy into an economic test without assuming probability calibration across model families.

What would settle it

Re-run the downstream evaluation with timestamped headlines aligned to prices immediately before publication and returns measured over 5-minute to 4-hour horizons, using market- or sector-adjusted abnormal returns; if any model's intraday rank ICs are positive and survive false-discovery-rate correction, the paper's next-session null would be overturned.

Watch

Extended reading notes

Core claim

The central claim is that QLoRA is effective for financial sentiment adaptation, while classification accuracy and economic predictability are distinct objectives: a model that reproduces human sentiment labels more accurately need not rank future returns better. Experiment 1 shows large supervised gains—Qwen2.5's macro-F1 rises from 0.7274 zero-shot to 0.8615 after QLoRA, while Mistral-7B and LLaMA3-8B reach 0.8771 and 0.8753—and that inverse-frequency class weighting does not help. Experiment 2 finds that all seven models produce positive but small one-day rank ICs between 0.0013 and 0.0143, that every model's IC turns negative at the two- or five-day horizon, and that no model-horizon test remains significant after multiple-testing correction. The paper interprets the positive multi-day long-only returns as market exposure rather than sentiment alpha, and treats the results as evidence of a gap between linguistic performance and economic usefulness.

Load-bearing premise

The economic null rests on the rule that every calendar-date headline signal is tested against returns from the first trading session strictly after the signal date, so if price discovery happens within the same session, the rank-IC and portfolio tests can be near zero even when sentiment carries real information.

Editorial extensions

If this is right

  • QLoRA can take a general 7-8B instruction-tuned LLM from roughly 0.73 to 0.86 macro-F1 on financial sentiment, so parameter-efficient adaptation is a practical route for specialized classifiers on limited hardware.
  • The best label-accuracy model is not the best economic-signal model: FinBERT, with the largest one-day rank IC, underperforms the QLoRA models on labels, so classifier leaderboards cannot substitute for return-based evaluation.
  • Any predictable component in these sentiment scores is concentrated in the first trading session and decays or reverses by days 2-5, consistent with fast price discovery.
  • The near-zero adjusted q-values (minimum 0.9622) imply that claims of financial sentiment alpha need multiple-testing-aware inference and cannot rest on one favorable horizon or one model.
  • Because long-only multi-day returns were broadly positive while long-short returns were negative, naive backtests that ignore market exposure can mistake beta for sentiment alpha.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Beyond the paper: the timing mismatch it identifies is directly testable—align each timestamped headline to returns starting minutes after publication over 5-minute to 4-hour horizons; if intraday rank ICs are positive and survive false-discovery-rate correction, the next-session null is an artifact of the alignment rule.
  • Beyond the paper: because the two Qwen variants produce nearly identical signals (score correlation 0.998), downstream behavior is driven by the QLoRA-adapted backbone rather than loss weighting; a return-aware training objective that maximizes rank IC would test whether the classification-to-predictability gap can be closed by design.
  • Beyond the paper: the sample's uneven coverage—usable headlines for only 72 constituents and small daily cross-sections—limits statistical power, so a multi-year, broader-universe replication could overturn the null even if the 2019 result stands.
  • Beyond the paper: the paper's own framing implies that a model trained to predict abnormal returns conditioned on expectations, rather than human sentiment labels, might show predictability that these label-trained classifiers miss.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 6 minor

Summary. This paper reports two experiments. Experiment 1 constructs a harmonized three-class benchmark from five financial text datasets and compares TF–IDF Naive Bayes, off-the-shelf FinBERT and Financial-RoBERTa, zero-shot Qwen2.5-7B, and QLoRA-adapted Qwen2.5-7B, LLaMA3-8B, and Mistral-7B, using a fixed train–validation–test split. Experiment 2 applies the seven probability-producing classifiers to 13,115 headline–stock observations for a fixed S&P 100 universe in 2019, converting model probabilities to continuous sentiment scores, aggregating by stock–date, and aligning them with next-session open returns over horizons of 1, 2, 3, and 5 days. The classification results show large gains from QLoRA relative to zero-shot Qwen (macro-F1 from 0.7274 to 0.8615), with Mistral the best at 0.8771. The downstream results show positive but small one-day rank ICs (largest 0.0143) and no test significant after Newey–West and FDR correction; portfolio returns are mostly weak or attributable to market exposure. The paper concludes there is a gap between classification accuracy and tradable cross-sectional signals.

Significance. If the results hold, the paper makes a useful empirical contribution: it provides a unified multi-dataset benchmark for financial sentiment classification with a clean same-backbone QLoRA comparison, and it demonstrates a careful evaluation protocol for linking sentiment scores to returns, including horizon-aware alignment, Newey–West inference, and FDR correction. The paper's explicit discussion of limitations (timing mismatch, small sample, contamination risk) is a strength. However, the strength of the economic conclusion currently exceeds what the evidence supports, because the null result is measured after the interval in which the signal may have already been traded.

major comments (3)
  1. [Section 3.5.2, Eq. (5)] The alignment rule maps every calendar-date signal to the first trading session strictly after date d, with entry at that session's open. The paper acknowledges in Section 6.1 that public information may be incorporated into large-cap equity prices within minutes or hours; the chosen alignment therefore measures the residual signal after the primary price adjustment has occurred. Because the abstract's central claim asserts a 'clear gap between classification accuracy and tradable cross-sectional signals', and this null result is consistent with a signal that expired before the measurement window, the gap is not documented by the current design. Please provide a sensitivity analysis with same-session or intraday alignment (e.g., using timestamped headlines for a subset) or revise the abstract and conclusion to characterize the result as 'no robust evidence of next-session predictability under the present alignment'.
  2. [Section 6.1] The downstream sample consists of one calendar year, 72 S&P 100 constituents with usable headlines, and 13,115 headline–stock observations over 253 dates. The paper notes that some daily cross-sections contain only a small number of stocks with fresh news, reducing power. With this limited power, the FDR-cleared null (minimum adjusted q-value 0.9622) is consistent both with a genuine absence of signal and with a weak signal that the test cannot detect. The paper should report a power analysis or an upper bound on the effect sizes that could be detected, and it should temper the 'clear gap' language in the abstract accordingly.
  3. [Section 6.1] The paper acknowledges that the pretrained models may have seen the 2019 Benzinga headlines in their training corpora, so the downstream evaluation is not strictly out-of-sample with respect to model weights. The abstract's phrase 'temporally separate' is correct but does not rule out contamination. Please attempt an empirical check (e.g., comparing model predictions on held-out vs. memorized headlines, or probing for exact-match recall) or explicitly qualify the 'out-of-sample' claim throughout the manuscript.
minor comments (6)
  1. [Abstract] The phrase 'clear gap between classification accuracy and tradable cross-sectional signals' is stronger than the paper's own conclusion in Section 6.1, which says the findings should be interpreted as evidence about the present experimental design rather than a general rejection of the economic value of sentiment. Please align the abstract with this qualification.
  2. [Section 2.1] In the sentence 'Words such asliability,capital, andtaxare frequently labeled as negative', there are missing spaces after 'as' and 'and'; please correct.
  3. [Table 1] The entry for Twitter Financial News Sentiment lists its year as '–' and gives no sample size; please complete the entry or note that the year is unknown.
  4. [Table 3] FinBERT and Financial-RoBERTa each appear in two rows ('Pretrained classification head' and 'Off-the-shelf sentiment checkpoint'), but the text says they are evaluated using the same publicly available checkpoints without fine-tuning. This duplication is confusing; please merge the rows or clarify the distinction.
  5. [Section 3.4] The classification comparisons are based on a single random seed (seed 42) without confidence intervals or repeated runs. The large zero-shot vs. QLoRA gap is clearly robust, but the small differences among Mistral, LLaMA3, and the two Qwen variants (macro-F1 0.8771 vs 0.8753 vs 0.8615) should not be over-interpreted. Reporting variance across several seeds would make the ranking claims more reliable.
  6. [General] The manuscript does not include code or data. Given the paper's benchmarking contribution, a reproducibility statement or link to the harmonized benchmark and evaluation scripts would strengthen the paper.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the benchmark, QLoRA comparison, and return tests are independent by construction.

full rationale

The paper's derivation chain is empirically self-contained rather than circular. Experiment 1 trains QLoRA models on a fixed training split and evaluates them on a held-out test split (Section 3.2.2), so the reported accuracy and macro-F1 gains are not fitted to the claimed conclusion. The downstream economic test in Experiment 2 uses frozen classifiers on a temporally separate 2019 Benzinga sample, and the continuous sentiment score s = p_pos - p_neg (Eq. 3) is defined purely from model probabilities; no return data enter model training or score construction. Returns appear only in the fixed alignment rule (Eq. 5) and in the rank-IC statistic (Eq. 6), which is a measurement rather than an estimated parameter. No parameter is fitted to the target return result, and no FDR threshold is tuned to manufacture the null. The paper's central claim that classification accuracy does not imply tradable cross-sectional signal is an empirical outcome, not a built-in consequence of any equation. The next-session alignment rule and the acknowledged possibility of intraday price discovery are validity threats that could explain the null, but they do not make the conclusion equivalent to the input: the null is not asserted by definition. The cited methods (QLoRA, LoRA, FinLoRA, AdaLoRA) are external prior work, and no load-bearing claim reduces to the author's own prior publication. Section 6.1 candidly lists timing mismatch, temporal contamination risk, and objective mismatch as limitations, which further confirms that the conclusions are not derived from those assumptions but are presented as contingent empirical findings.

Assumptions & free parameters 9 free parameters · 5 assumptions · 0 invented entities

The central claims rest on standard supervised evaluation and standard return-predictability statistics. The nonstandard commitments are the label harmonization, the next-session timing alignment, the 2019 Benzinga sample as the economic test bed, and the assumption that post-2019 pretraining does not materially contaminate that sample. All of these are acknowledged as limitations in the paper.

free parameters (9)
  • Random seed = 42
    Controls stratified splits and training randomness; single-seed results have no variance estimate.
  • LoRA rank r = 16
    Common adapter rank chosen for all backbones; not swept.
  • LoRA scaling alpha = 32
    Set to 2*r, standard QLoRA choice.
  • Learning rate = 2e-5
    Optimizer setting from Table 5.
  • Epochs = 3
    Training duration from Table 5.
  • Gradient accumulation steps = 32
    Effective batch size of 32, chosen in Table 5.
  • Portfolio long-short quantile = 15%
    Top and bottom 15% ranking tails in Section 3.5.4; changes point estimates.
  • Return horizons = 1, 2, 3, 5 trading days
    Horizon set in Eq. (5); the sign pattern of ICs is horizon-dependent.
  • Sentiment score mapping = s = p_pos - p_neg
    Ad hoc continuous score defined in Eq. (3); neutral has zero, no calibration across models.
assumptions (5)
  • standard math Spearman rank correlation, Newey-West HAC standard errors, and Benjamini-Hochberg FDR are valid for the overlapping multi-day return tests.
    Used in Section 3.5.3 for inference on the 28 model-horizon IC tests.
  • domain assumption The Benzinga 2019 headline set for the available S&P 100 constituents is a sufficient out-of-sample sample for the economic tests.
    Section 3.5.1; the paper acknowledges coverage of only 72 constituents and uneven news frequency.
  • domain assumption Calendar-date headline signals can be aligned to the first trading session strictly after the signal date without intraday timestamps.
    Section 3.5.2 and Eq. (5); this timing choice is load-bearing for the null result.
  • domain assumption Native labels from FPB, FOMC, SEntFiN, Twitter, and NASDAQ can be mapped into one common three-class sentiment space.
    Section 3.2.1; the paper states the original annotation tasks are not semantically identical.
  • domain assumption Post-2019 pretraining of the LLMs does not materially contaminate the 2019 Benzinga evaluation.
    Section 6.1 says this possibility cannot be excluded, so the 'temporally separate' framing depends on it.

how reviews work

0 comments
Cite this review

Pith. "Pith review of From Financial Sentiment Classification to Return Predictability: A QLoRA Benchmark of Large Language Models." pith.science (2026). https://pith.science/paper/SRHIM4GM

@misc{pith2026260804200,
  author       = {Pith},
  title        = {Pith review of: From Financial Sentiment Classification to Return Predictability: A QLoRA Benchmark of Large Language Models},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/SRHIM4GM}},
  note         = {Machine review of arXiv:2608.04200}
}
read the original abstract

Financial sentiment classifiers are commonly evaluated against human labels, but strong linguistic performance does not necessarily imply economically useful return predictability. This study separates these questions through two experiments. First, we construct a unified three-class benchmark from five financial text datasets and compare TF--IDF Naive Bayes, off-the-shelf FinBERT and Financial-RoBERTa encoders, zero-shot Qwen2.5-7B, and QLoRA-adapted Qwen2.5-7B, LLaMA3-8B, and Mistral-7B models. Mistral-7B achieves the best test accuracy (0.8840) and macro-F1 (0.8771), while QLoRA raises Qwen2.5's macro-F1 from 0.7274 to 0.8615. An inverse-frequency class-weighted loss does not improve Qwen2.5. Second, we evaluate economic validity on a temporally separate 2019 Benzinga sample containing 10,637 unique headlines and 13,115 headline--stock observations for a fixed S\&P~100 universe. Model probabilities are converted into continuous sentiment scores, aggregated by stock and signal date, and aligned with next-session returns over one-, two-, three-, and five-day horizons. All seven downstream models produce positive but small mean rank information coefficients at the one-day horizon; the largest is 0.0143 for FinBERT. None of the 28 model--horizon tests remains significant after Newey--West inference and false-discovery-rate correction. Portfolio results likewise fail to establish a robust advantage for the best-performing classifiers. The findings show that QLoRA is effective for financial sentiment adaptation, while also documenting a clear gap between classification accuracy and tradable cross-sectional signals.

Figures

Figures reproduced from arXiv: 2608.04200 by the authors.

Figure 1
Figure 1. Experiment 2 pipeline. The same fresh-news stock–date sample is used for model-specific Rank IC tests and equal [PITH_FULL_IMAGE:figures/full_fig_p009_1.png] view at source ↗
Figure 2
Figure 2. Predicted sentiment-label shares across seven models on the unlabeled 2019 Benzinga S&P 100 unique-news sample [PITH_FULL_IMAGE:figures/full_fig_p012_2.png] view at source ↗
Figure 3
Figure 3. Model-specific distributions of continuous headline sentiment scores on the Benzinga sample. The score is defined as [PITH_FULL_IMAGE:figures/full_fig_p013_3.png] view at source ↗
Figures from the paper (2 more)
Figure 4
Figure 4. Figure 4: Mean prediction confidence of each model on the Benzinga sample, measured as the maximum predicted probability [PITH_FULL_IMAGE:figures/full_fig_p014_4.png]
Figure 5
Figure 5. Figure 5: Multi-horizon downstream evaluation. Panel (a) reports the mean daily cross-sectional Spearman rank IC between [PITH_FULL_IMAGE:figures/full_fig_p014_5.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

21 extracted references · 7 canonical work pages

  1. [1]

    Dogu Araci. 2019. FinBERT: Financial Sentiment Analysis with Pre-trained Language Models. arXiv:1908.10063 [cs.CL] https://arxiv.org/abs/1908.10063

    arXiv 2019

  2. [2]

    Tim Dettmers, Artidoro Pagnoni, Ari Holtzman, and Luke Zettlemoyer. 2023. QLoRA: Efficient Finetuning of Quantized LLMs. arXiv:2305.14314 [cs.LG] https: //arxiv.org/abs/2305.14314

    arXiv 2023

  3. [3]

    Zeyu Han, Chao Gao, Jinyang Liu, Jeff Zhang, and Sai Qian Zhang. 2024. Parameter-Efficient Fine-Tuning for Large Models: A Comprehensive Survey. arXiv:2403.14608 [cs.LG] https://arxiv.org/abs/2403.14608

    arXiv 2024

  4. [4]

    Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen

    Edward J. Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. 2021. LoRA: Low-Rank Adaptation of Large Language Models. arXiv:2106.09685 [cs.CL] https://arxiv.org/abs/2106.09685

    arXiv 2021

  5. [5]

    Dieuwke Hupkes, Mario Giulianelli, Verna Dankers, Mikel Artetxe, Yanai Elazar, Tiago Pimentel, Christos Christodoulopoulos, Karim Lasri, Naomi Saphra, Ara- bella Sinclair, Dennis Ulmer, Florian Schottmann, Khuyagbaatar Batsuren, Kaiser Sun, Koustuv Sinha, Leila Khalatbari, Maria Ryskina, Rita Frieske, Ryan Cotterell, and Zhijing Jin. 2023. A taxonomy and ...

    2023

  6. [6]

    Aakanksha Jadhav and Vishal Mirza. 2025. Large language models in equity markets: Applications, techniques, and insights.Frontiers in Artificial Intelligence 8 (2025), 1608365

    work page 2025

  7. [7]

    Aikaterini Karanikola, Gregory Davrazos, Charalampos M Liapis, and Sotiris Kotsiantis. 2023. Financial sentiment analysis: Classic methods vs. deep learning models.Intelligent Decision Technologies17, 4 (2023), 893–915

    work page 2023

  8. [8]

    Yixiao Li, Yifan Yu, Chen Liang, Pengcheng He, Nikos Karampatziakis, Weizhu Chen, and Tuo Zhao. 2023. LoftQ: LoRA-Fine-Tuning-Aware Quantization for Large Language Models. arXiv:2310.08659 [cs.CL] https://arxiv.org/abs/2310. 08659

    arXiv 2023

Show all 21 references
  1. [9]

    TIM LOUGHRAN and BILL MCDONALD. 2011. When Is a Liability Not a Liability? Textual Analysis, Dictionaries, and 10-Ks.The Journal of Finance66, 1 (2011), 35–65. arXiv:https://onlinelibrary.wiley.com/doi/pdf/10.1111/j.1540- 6261.2010.01625.x doi:10.1111/j.1540-6261.2010.01625.x

    2011

  2. [10]

    Macedo Maia, Siegfried Handschuh, Andre Freitas, Brian Davis, Ross McDermott, Manel Zarrouk, and Alexandra Balahur. 2018. WWW’18 Open Challenge: Finan- cial Opinion Mining and Question Answering.WWW ’18: Companion Proceedings of the The Web Conference 2018, 1941–1942. doi:10.1...

    2018

  3. [11]

    P. Malo, A. Sinha, P. Korhonen, J. Wallenius, and P. Takala. 2014. Good debt or bad debt: Detecting semantic orientations in economic texts.Journal of the Association for Information Science and Technology65 (2014)

    2014

  4. [12]

    Paul Wang, Huidi Lu, W

    Laiyi Meng, P. Paul Wang, Huidi Lu, W. Yuna Yang, and Suhas Vijayakumar. 2026. Entity-specific text classification with large language models: two applications on financial and brand news.China Accounting and Finance Review(05 2026), 1–

    2026

  5. [13]

    arXiv:https://www.emerald.com/cafr/article-pdf/doi/10.1108/CAFR-12-2025- 0269/11622137/cafr-12-2025-0269en.pdf doi:10.1108/CAFR-12-2025-0269

    2025 doi

  6. [14]

    Agam Shah, Suvan Paturi, and Sudheer Chava. 2023. Trillion Dollar Words: A New Financial Dataset, Task & Market Analysis. InProceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). Association for Computational Linguistic...

    2023 doi

  7. [15]

    Numair Shaikh, Jayesh Patil, and Sheetal Sonawane. 2023. Query-based sum- marization and sentiment analysis for Indian financial text by leveraging dense passage retriever, RoBERTa, and FinBERT. InProceedings of the 20th International Conference on Natural Language Processing ...

    2023

  8. [16]

    Ankur Sinha, Satishwar Kedas, Rishu Kumar, and Pekka Malo. 2022. <scp>SEntFiN</scp> 1.0: <scp>Entity-aware</scp> sentiment analysis for fi- nancial news.Journal of the Association for Information Science and Technology 73, 9 (March 2022), 1314–1335. doi:10.1002/asi.24634

    2022 doi

  9. [17]

    Dannong Wang, Daniel Kim, Bo Jin, Xingjian Zhao, Tianfan Fu, Steve Yang, and Xiao-Yang Liu. 2025. FinLoRA: Finetuning Quantized Financial Large Language Models Using Low-Rank Adaptation. arXiv:2412.11378 [cs.LG] https://arxiv.org/ abs/2412.11378

    2025 arXiv

  10. [18]

    Shijie Wu, Ozan Irsoy, Steven Lu, Vadim Dabravolski, Mark Dredze, Sebastian Gehrmann, Prabhanjan Kambadur, David Rosenberg, and Gideon Mann. 2023. BloombergGPT: A Large Language Model for Finance. arXiv:2303.17564 [cs.LG] https://arxiv.org/abs/2303.17564

    2023 arXiv

  11. [19]

    Qianqian Xie, Weiguang Han, Zhengyu Chen, Ruoyu Xiang, Xiao Zhang, Yueru He, Mengxi Xiao, Dong Li, Yongfu Dai, Duanyu Feng, et al . 2024. Finben: A holistic financial benchmark for large language models.Advances in neural information processing systems37 (2024), 95716–95743

    2024

  12. [20]

    Hongyang Yang, Xiao-Yang Liu, and Christina Dan Wang. 2025. FinGPT: Open- Source Financial Large Language Models. arXiv:2306.06031 [q-fin.ST] https: //arxiv.org/abs/2306.06031

    2025

  13. [21]

    Qingru Zhang, Minshuo Chen, Alexander Bukharin, Nikos Karampatziakis, Pengcheng He, Yu Cheng, Weizhu Chen, and Tuo Zhao. 2023. AdaLoRA: Adaptive Budget Allocation for Parameter-Efficient Fine-Tuning. arXiv:2303.10512 [cs.CL] https://arxiv.org/abs/2303.10512

    2023 arXiv

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.