arXiv cs.CL· Alina Khaybullina·· 2 天前AI 评分22
LLM 财务预测中的动量、评论语气与盈利意外
Momentum, Commentary Tone, and Earnings Surprise in LLM Financial Forecasts
AI 导读
研究用合成财报测试 LLM 财务预测的输入依赖：对四个模型家族六个指令/对话 checkpoint（0.49-7.61B 基础参数）各输入 2,016 条提示，主要 EPS 对比跨度达 -0.19 至 86.80 个百分点。六个模型中有五个的带符号语气对比幅度超过 EPS，动量也在五个模型中超过 EPS，自然词数值诊断准确率为 51.0% 至 95.8%。
正文 · 原文
Abstract:Which inputs drive a language model's financial forecast? We construct synthetic earnings announcements that independently reverse 20-day momentum, non-informational commentary tone, and whether fixed actual EPS beats a manipulated consensus. Six instruction/chat checkpoint conversions from four families (0.49-7.61B base parameters) receive 2,016 prompts each under a direct-answer Q8_0 protocol. The design includes 24 numerical blocks, six commentary pairs, profits and losses, rising and falling annual EPS, and two surprise sizes. Both A/B answer assignments and U/D cover all 384 primary scenarios per conversion. We estimate matched changes in conditional answer-prefix scores, not calibrated market probabilities. Across these checkpoints, the primary EPS contrast ranges from -0.19 to 86.80 percentage points. The magnitude of the primary signed tone contrast exceeds that of EPS in 5 of six models, and momentum exceeds it in 5. Natural-word numerical diagnostic accuracy ranges from 51.0 to 95.8 percent. Answer-encoding, wording, context, and conversion controls qualify these results. The experiment measures observable input reliance and numerical recognition; it establishes neither an internal reasoning mechanism nor realized return predictability. Public benchmark prompts, recorded scores, and an interactive recorded-results explorer accompany the study.
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); General Finance (q-fin.GN) |
| Cite as: | arXiv:2610.22043 [cs.LG] |
| (or arXiv:2610.22043v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.22043 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Alina Khaybullina [view email]
[v1]
Thu, 15 Oct 2026 15:14:20 UTC (1,533 KB)
来源：arXiv cs.CL · arxiv.org