Deletion Tuning matches full fine-tuning by zeroing under 0.05% of weights
A binary mask on gate projections stays near base prior-task accuracy while FFT, LoRA, and random sparse tuning lose up to 5.4 points.
Stories per day · 30 days
1 RSS
Relatedcontinual-learning1distillation1fine-tuning1lora1mixture-of-experts1qwen31reasoning1sparse-training1
5 stories, newest first · one card per story
Stories tagged #llm-post-training
Calibrated distillation consolidates specialists in dynamic MoE models
On Turing-20B-A2B the method scores 0.9988 across four domains and beats Qwen3-8B on math and code.
DiffGate: Difficulty-Gated Teacher Guidance for On-Policy Distillation
DiffGate gates bounded teacher guidance to failed rollouts, lifting Qwen3 pass rates over matched GRPO.
Learning to Steer, Steering to See: Unveiling the Geometry of RLVR in Large Language Models via Trainable Vectors
Vector steering finds a low-dimensional manifold behind RLVR gains and a stabilizer that preserves them.
PACT: From Credit Assignment to Critic Alignment
PACT realigns actor-critic training and beats GRPO and PPO on math reasoning and SWE-bench Verified.