Self-Correction: GPT-4 falls 95.5%→91.5% on GSM8K
Huang et al. (ICLR 2024): intrinsic self-correction drops GPT-4 from 95.5% to 91.5% on GSM8K—ledger agents need external validators, not self-review.
#machine-learning
Machine Learning
Machine learning techniques for financial data analysis and automation
Tree of Thoughts: Deliberate Problem Solving with LLM Search
Tree of Thoughts (ToT) achieves 74% on Game of 24 vs 4% for standard GPT-4 CoT by organizing LLM reasoning into a branching search tree with pruning and backtracking — with direct implications for multi-step financial classification and tax optimization in Beancount workflows.
CRITIC: Why LLM Self-Correction Requires External Tool Feedback
CRITIC (ICLR 2024) achieves 7.7 F1 gains on open-domain QA and a 79.2% toxicity reduction by grounding LLM revision in external tool signals — a verify-then-correct loop that maps directly onto write-back safety for Beancount finance agents.
Reflexion: Language Agents That Learn from Mistakes Without Retraining
Reflexion (NeurIPS 2023) lets LLM agents improve by storing verbal post-mortems in an episodic buffer — no weight updates required. It reaches 91% on HumanEval with GPT-4 but fails on WebShop, revealing a structural constraint: verbal reinforcement only works when the evaluator produces a crisp, actionable signal. Here is what that means for building a self-correcting Beancount ledger agent.
Self-Consistency: Majority-Vote Sampling Improves Chain-of-Thought Accuracy
Self-consistency replaces greedy chain-of-thought decoding with a majority vote over N sampled reasoning paths — raising GPT-3 accuracy on GSM8K by 17.9 percentage points with no extra training — and applies directly to multi-step financial calculations where a single model decode is unreliable.
PAL: +38pp on hard math by handing arithmetic to Python
PAL gains +38.1pp over chain-of-thought on GSM-hard by running Python for arithmetic—the right split for reliable Beancount ledger calculations.
Constitutional AI for Accounting Agents: RLAIF, Policy Rules, and Goodharting Risks
Anthropic's Constitutional AI paper (Bai et al., 2022) trains LLMs to follow rules using AI-generated feedback rather than human harm labels. This research log examines how the RLAIF critique-revise-preference pipeline maps onto write-back safety for autonomous Beancount ledger agents — and what Goodharting, calibration failures, and dual-use risks look like when the "constitution" is a chart of accounts instead of an ethics ruleset.
Chain-of-Thought Prompting: Precision-Recall Trade-offs for Finance AI
A close reading of Wei et al.'s 2022 Chain-of-Thought paper and what it means for finance AI — why CoT raises precision but may cut recall on rare-event detection, why the scale threshold matters for production agents, and what a finance team building on LLMs should watch out for.
PHANTOM (NeurIPS 2025): Measuring LLM Hallucination Detection in Financial Documents
PHANTOM (NeurIPS 2025) is the first benchmark to measure LLM hallucination detection on real SEC filings across context lengths up to 30,000 tokens. Qwen3-30B-A3B-Thinking leads with F1=0.882; 7B models score near random guessing — with direct implications for autonomous accounting agents.
FinMaster Benchmark: Why LLMs Score 96% on Financial Literacy but 3% on Statement Generation
FinMaster (arXiv:2505.13533) benchmarks o3-mini, Claude 3.7 Sonnet, and DeepSeek-V3 across 183 financial tasks—revealing that models score 96% on financial literacy but collapse to 3% on statement generation, with multi-step consulting tasks losing 21 accuracy points from error propagation.
ReAct: Synergizing Reasoning and Acting in Language Models
ReAct (Yao et al., ICLR 2023) interleaves chain-of-thought reasoning with tool actions in a single trajectory, outperforming pure CoT on fact verification and imitation learning on embodied tasks by 34 percentage points. This analysis covers the paper's failure modes — search-induced distraction and compounding errors — and what they mean for autonomous agents writing back to Beancount ledgers.
Toolformer: Self-Supervised Tool Use and Its Limits for Finance AI
A close reading of Toolformer (Meta AI, NeurIPS 2023): how perplexity-filtered self-supervised training teaches a 6.7B-parameter model to call external APIs, where it outperforms GPT-3 175B on arithmetic benchmarks, and why its single-step architecture cannot support the chained tool calls required for structured ledger operations.