Extracting Long Conversation States with LoRA
Small-Inevitable6185 · reddit · 2026-07-12
The author is working on a LoRA project for long conversation continuity, aiming not to summarize the entire chat, but to extract a structured "conversation state" from each chunk and merge them into context for subsequent generation.
The current workflow roughly involves:
- Splitting long conversations into fixed windows
- Annotating semantic states for each window
- Fine-tuning a small model with LoRA
- Merging outputs from all windows to form a state representation for continued conversation
- Generating a continuation prompt
Real-world engineering data was used, including GitHub Issues, GitHub Discussions, Reddit engineering threads, and long AI dev conversations. Several common clusters were identified: context/memory management, state persistence, reliability, provider compatibility, agent orchestration, long debugging sessions, and architecture discussions.
The author is currently debating whether to invest more in datasets, coverage, and annotation evaluation, or to drop fine-tuning entirely in favor of a stronger base model with prompting.
More from coding & agent
- Coding agents are heading toward an AI-writes, AI-reviews, human-approves workflow — aftahi_ai · 2026-07-22
- oMLX 0.5.2 adds Mac menu-bar stats, low-bit decode kernels, and faster downloads — awnihannun · 2026-07-22
- GitHub review bot hits its PR limit and forces a 39-minute cooldown — DanielLockyer · 2026-07-22
- Max reasoning effort appears to be mobile-only in Codex Remote, not desktop — GabGarrett · 2026-07-22
- A Reddit demo argues online stores should expose carts and pricing through MCP — gelembjuk · 2026-07-22
- Open-source AI SDK provider routes Vercel apps through a local Codex subscription — lgrammel · 2026-07-22