Parsing Messy Language into DSL: Stability Challenges with DeepSeek V4 Flash
Ok_Pop3337 · reddit · 2026-08-25
A developer shares experience building a conversational finance tool that converts natural language to typed financial operations using DeepSeek V4 Flash. While deterministic validation works safely, the LLM struggles with semantic stability: dropping plausible interpretations, merging distinct operations, and hallucinating fields like dates. Attempts using stricter schemas, larger prompts, high reasoning effort, and separate judging steps yielded limited improvements. The author seeks advice on retrieval strategies, intent classification, intermediate representations, constrained decoding, specialist agents, and evaluation methods.
More from coding & agent
- Dev report: Running agents and UI on a 64GB M4 Mini — doodlestein · 2026-08-26
- Openwiki v0.4.0: Introduces Forgetting Mechanism to Handle Wiki Updates — LangChain · 2026-08-26
- Issue: Xcode failing to sign in to Apple Accounts on Golden Gate VM — brianmichel · 2026-08-26
- YouArt launches MCP integration for visual generation within Claude and Codex — SimplyAnnisa · 2026-08-26
- You Can't Claim Models Are Good Enough to Skip Code Review: HamelHusain RT — HamelHusain · 2026-08-26
- NeurIPS workshop CFP: Interpreting Agent Behavior — mdredze · 2026-08-26