Influcoder: Distilling gradient influence rankings into an encoder for scalable data attribution
Influcoder is a proposed method for scalable data attribution in LLM training, distilling decoder-based gradient influence rankings into a compact encoder representation. The approach targets the practical bottleneck of influence function methods — their high computational cost and storage requirements — making them viable for large-scale dataset curation. The work is relevant to training data quality filtering and identifying sources of undesirable model behavior such as toxicity.
Related guides (2)
Related events (8)
GGRO: Gradient-Guided Reward Optimization for inference-time LLM alignment
Researchers introduce Gradient-Guided Reward Optimization (GGRO), an inference-time alignment method that uses gradient signals from a reward model to inject 'nudging tokens' at high-uncertainty decoding steps, rather than relying on sampling-intensive re-ranking approaches like Best-of-N. The method monitors token-level entropy to detect distribution drift and steers generation trajectories directly, claiming improved robustness to reward hacking with minimal computational overhead. Experiments show gains across safety, helpfulness, and reasoning benchmarks compared to standard inference-time alignment baselines.
Interpretability-based pipeline for auditing and shaping post-training learning signals
Researchers introduce a data-centric post-training pipeline that applies interpretability methods to preference datasets before optimization, surfacing latent concepts that separate preferred from dispreferred generations. The approach unifies several interpretability-based training protocols as feature or data interventions that shape reward signals. Empirically, the pipeline diagnoses undesirable signals such as sycophancy and over-stylization, mitigates off-target learning, and can amplify desired properties like safety behaviors and model personality. The work reframes post-training from opaque scalar reward optimization into an auditable, concept-level sculpting process.
Optimization story: Bloom inference
This Hugging Face blog post documents practical inference optimization techniques applied to the BLOOM large language model. It covers strategies for reducing latency and memory footprint during deployment, likely including quantization, tensor parallelism, and batching approaches. The post serves as a technical case study for serving very large open-weights models efficiently.
STARE: Token-level advantage reweighting to prevent entropy collapse in GRPO-style RL training
Researchers introduce STARE, a method addressing policy entropy collapse in GRPO-style reinforcement learning from verifiable rewards (RLVR) for LLM post-training. Through first-order gradient analysis, they identify a token-level credit assignment mismatch and propose selectively reweighting advantages for entropy-critical tokens using batch-internal surprisal quantiles plus a closed-loop entropy gate. Evaluated across 1.5B–32B models on short/long chain-of-thought and multi-turn tool use tasks, STARE outperforms DAPO and other baselines by 4–8% on AIME24/25 while sustaining stable training over thousands of steps.
Apple researchers propose Feature Auto-Encoder to speed diffusion training via compressed DINOv2 embeddings
Researchers at Apple introduced Feature Auto-Encoder (FAE), a latent diffusion image generator that compresses DINOv2 vision encoder embeddings before learning to denoise them, then expands them back for decoding. The approach achieves comparable image quality to state-of-the-art diffusion models while training roughly 7x faster on ImageNet class-conditional generation. The key insight is that shrinking semantically rich vision embeddings reduces compute during diffusion training without sacrificing the representational benefits of large pretrained encoders.
Forecasting Downstream LLM Performance With Token-Level Proxy Metrics
Researchers propose proxy metrics constructed from token-level statistics (entropy, top-k accuracy, expert token rank) drawn from a candidate model's next-token distribution over expert-written solutions, as a cheaper and more reliable alternative to cross-entropy loss or direct downstream evaluation. Across three settings—cross-family model selection, pretraining data selection, and training-time forecasting—the proxies consistently outperform baselines, achieving mean Spearman Rho of 0.81 vs. 0.36 for cross-entropy loss on model ranking, and reducing compute for data selection by roughly 10,000×. The method enables downstream performance extrapolation across an 18× compute horizon with roughly half the error of existing alternatives, suggesting expert trajectories are broadly useful signals throughout the model development lifecycle.
Systematic study reveals effectiveness-fluency trade-offs in LLM conditioning methods
A new arXiv paper systematically evaluates a range of LLM conditioning methods across both concept injection and removal scenarios, finding that efficient steering methods often degrade fluency significantly. A key finding is that activation steering is substantially less effective on instruction-tuned models than on base models, a previously overlooked interaction. Simple prompting and supervised fine-tuning work for concept injection but not removal, and cheap textual metrics are found to correlate well with expensive LLM-as-judge evaluations.
AGDO: Attention-guided denoising and optimization framework improves diffusion language model reasoning
Researchers propose AGDO, a framework that replaces random masking in diffusion large language models (dLLMs) with attention-guided denoising order and token weighting during fine-tuning and reinforcement learning. The work is motivated by an empirical finding that tokens with stronger attention to unmasked context are more stable and critical for reasoning. Experiments on math and coding benchmarks show AGDO outperforms existing post-training methods for dLLMs, advancing the case for attention-aware training in parallel-decoding language models.