Issues

• Group chats produce low-signal training data (all sources)

  enhancement
  NotYuSheng/Doppelganger#46 · Filter by author NotYuSheng opened on Sep 13, 2026

• Explore: domain-specific tuning beyond general writing style

  exploration
  NotYuSheng/Doppelganger#45 · Filter by author NotYuSheng opened on Jun 25, 2026

• Explore: safety against extraction — people probing the model to scrape personal info

  exploration
  NotYuSheng/Doppelganger#44 · Filter by author NotYuSheng opened on Jun 25, 2026

• Explore: interpretability — "what did it learn about me?"

  exploration
  NotYuSheng/Doppelganger#38 · Filter by author NotYuSheng opened on Jun 23, 2026

• Explore: prompt optimization (DSPy) for the persona prompt

  exploration
  NotYuSheng/Doppelganger#37 · Filter by author NotYuSheng opened on Jun 23, 2026

• Explore: activation steering / control vectors (inference-time tone)

  exploration
  NotYuSheng/Doppelganger#36 · Filter by author NotYuSheng opened on Jun 23, 2026

• Explore: memorization audits, canaries, watermarking, federated

  exploration
  NotYuSheng/Doppelganger#35 · Filter by author NotYuSheng opened on Jun 23, 2026

• Explore: machine unlearning (right to be forgotten)

  exploration
  NotYuSheng/Doppelganger#34 · Filter by author NotYuSheng opened on Jun 23, 2026

• Explore: differential privacy training (DP-SGD)

  exploration
  NotYuSheng/Doppelganger#33 · Filter by author NotYuSheng opened on Jun 23, 2026

• Explore: proactive / initiative modeling

  exploration
  NotYuSheng/Doppelganger#32 · Filter by author NotYuSheng opened on Jun 23, 2026

• Explore: multi-agent & self-play (doppelgangers talking)

  exploration
  NotYuSheng/Doppelganger#31 · Filter by author NotYuSheng opened on Jun 23, 2026

• Explore: non-text behavior (stickers / emoji / memes)

  exploration
  NotYuSheng/Doppelganger#30 · Filter by author NotYuSheng opened on Jun 23, 2026

• Explore: multimodal voice (TTS voice cloning + STT)

  exploration
  NotYuSheng/Doppelganger#29 · Filter by author NotYuSheng opened on Jun 23, 2026

• Explore: style embeddings / user-conditioning

  exploration
  NotYuSheng/Doppelganger#28 · Filter by author NotYuSheng opened on Jun 23, 2026

• Explore: relationship/knowledge graph from chats

  exploration
  NotYuSheng/Doppelganger#27 · Filter by author NotYuSheng opened on Jun 23, 2026

• Explore: long-term memory architecture (episodic/semantic + reflection)

  exploration
  NotYuSheng/Doppelganger#26 · Filter by author NotYuSheng opened on Jun 23, 2026

• Explore: PEFT method comparison (LoRA vs DoRA / IA3 / prefix-tuning)

  exploration
  NotYuSheng/Doppelganger#25 · Filter by author NotYuSheng opened on Jun 23, 2026

• Explore: knowledge distillation -> on-device doppelganger

  exploration
  NotYuSheng/Doppelganger#24 · Filter by author NotYuSheng opened on Jun 23, 2026

• Explore: multi-LoRA personas & model merging

  exploration
  NotYuSheng/Doppelganger#23 · Filter by author NotYuSheng opened on Jun 23, 2026

• Explore: synthetic data generation / self-instruct (sparse-history augmentation)

  exploration
  NotYuSheng/Doppelganger#22 · Filter by author NotYuSheng opened on Jun 23, 2026

• Explore: MCP integration (tools & context via Model Context Protocol)

  exploration
  NotYuSheng/Doppelganger#20 · Filter by author NotYuSheng opened on Jun 23, 2026

• Explore: agentic doppelganger — tool use and actions

  exploration
  NotYuSheng/Doppelganger#19 · Filter by author NotYuSheng opened on Jun 23, 2026

• Explore: continual learning — keep the doppelganger current

  exploration
  NotYuSheng/Doppelganger#18 · Filter by author NotYuSheng opened on Jun 23, 2026

• Explore: evaluation — "does this actually sound like me?"

  exploration
  NotYuSheng/Doppelganger#17 · Filter by author NotYuSheng opened on Jun 23, 2026

• Explore: RAG / retrieval-augmented memory for the doppelganger

  exploration
  NotYuSheng/Doppelganger#16 · Filter by author NotYuSheng opened on Jun 23, 2026