fine-tuning
30 posts tagged fine-tuning.
- Fine-tuning 3D generation models is an asset pipeline problem
- Arete-27B tests values as a post-training layer
- A2ODA is a small optimizer idea with a practical training lesson
- Fine-tuning can break alignment without harmful data
- Sonar-MoE cuts MoE fine-tuning memory by moving optimizer steps into backward
- TopKLoRA turns a poisoned fine-tune into a circuit you can cut out
- A 4B Qwen fine-tune points at local 3D scene generation
- WeAct Turns LoRA Adapters Into Something an Agent Can Pick Mid-Task
- Reasoning distillation works better when the trace fits the student
- When Fine-Tuning for One Behavior Drags Another Along
- HiRe-MoMa fine-tunes robots by correcting policies instead of replacing them
- Back-translation beats more data in low-resource translation
- SGUID says most skills are not worth distilling
- The missing model is becoming a product requirement
- Teaching a Reasoning Model to Know When It's Sure Cuts Its Token Bill
- Translation Fine-Tuning Breaks the Controls General Evals Miss
- ComPO and the Case Against Optimizing the Loss You Wrote Down
- Distillation needs calibration when the teacher is biased
- What Fyxer's AI inbox assistant gets right about trust
- The Hallucination Detector That Doesn't Transfer to Your Domain
- The SFT-RL Split Has a Wide Safe Zone, and You Can Find It Cheap
- What LoRA Rank Actually Buys You, According to the Attention Math
- Reasoning fine-tunes can move the safety vector
- What the Summer 2026 Open Model Data Actually Shows
- ROPD treats poisoned fine-tunes as a distribution problem
- κ-LoRA makes LoRA tuning selective instead of uniform
- LoRA patches may outlive the base model update
- PPL-Factory makes the case for smaller fine-tuning sets
- CRAFT Turns Rubrics Into a Fine-Tuning Map
- When the Task Matches the Objective: What MTO Says About Fine-Tuning Small Models