Claims Browser

Successive Capacity Growth (SCG) achieves 20.3% improvement in prediction loss over fixed small baseline with 56 times greater parameter efficiency on multi-object dynamics tasks

"On a 60-dimensional multi-object dynamics task, SCG naturally triggers depth expansion, improving prediction loss by 20.3% over the fixed small baseline with 56 times greater parameter efficiency than scaling to the fixed large model"

SCG achieves 23% improvement over fixed large models on 2D navigation tasks with a single width expansion

"on a 2D navigation task, a single width expansion yields even an 23% improvement over the fixed large model"

The Sketched Isotropic Gaussian Regularizer (SIGReg) prevents representation collapse by ensuring all learned semantic dimensions remain statistically independent and aligned with the predictive objective

"The Sketched Isotropic Gaussian Regularizer (SIGReg) ensures that all learned semantic dimensions remain statistically independent and aligned with the predictive objective, preventing collapse even as the architecture grows"

SCG's adaptive encoder matches or exceeds fixed small baseline across all tested environments with zero false-positive expansions and bit-exact function preservation

"Across all three tested environments of increasing complexity, the adaptive encoder matches or exceeds the fixed small baseline, with zero false-positive expansions and bit-exact function preservation (ratio = 1.0, absolute difference = 0.0)"

Training Llama-3.2-3B costs over $1.5M at small scale

"Even at a small scale, training Llama-3.2-3B costs over \$1.5M"

Reproducing SmolLM3-3B requires over $700K in training costs

"reproducing SmolLM3-3B needs over \$700K"

Puro-2B can be trained for less than $6.9K while approaching Qwen2.5-1.5B performance

"Our best model is trained at a compute cost of less than \$6.9K and approaches Qwen2.5-1.5B performance under our evaluation protocol"

FP8 precision training on consumer-grade RTX 5090 GPUs enables dramatic cost reduction for language model pretraining

"we train a collection of Puro-2B models from scratch on up to 1.4 trillion tokens with FP8 precision on consumer-grade RTX 5090 GPUs"

The D2C-Routing detector system achieves 0.8603 four-way Avg TPR@1%FPR on mixed-origin text detection, outperforming RACE-local by 6.5 points

"our disclosed D2C-Routing-based detector system reaches 0.8603 four-way Avg TPR@1%FPR, 6.5 points above the same-split RACE-local rerun"

Distinguishing AI-content/human-expression from fully AI-generated text is the hardest classification boundary in mixed-origin detection

"error analysis shows that distinguishing AI-content/human-expression from fully AI-generated text remains the hardest boundary"

Generative AI is transforming how people access information and challenging traditional advertising mechanisms

"Generative AI is transforming how people access information, challenging traditional advertising mechanisms built around predefined slots."

The Latent Advertiser Mixture Auction (LAMA) mechanism satisfies Markov DSIC and IR properties while achieving near-optimal KL-regularized welfare

"We show that LAMA satisfies Markov DSIC and IR, and achieves near-optimal KL-regularized welfare."

LAMA improves platform welfare and revenue while maintaining user-facing response quality in commercial search applications

"Proof-of-concept experiments on real-world commercial-search query splits show that LAMA improves platform welfare and revenue while maintaining user-facing response quality"

LLMs show increasingly poor performance as input size approaches realistic corporate communication scales of 230,000+ documents

"We evaluate five LLMs on CB, revealing increasingly poor performance as input size approaches realistic scales."

LLMs are increasingly capable of answering complex questions about enterprise-scale document collections

"LLMs are increasingly able to answer complex questions about enterprise-scale document collections."

NPO achieves comparable or better performance than GEPA with fewer rollouts

"NPO achieves comparable or better performance than GEPA with fewer rollouts"

Prompt optimization can deliver performance gains comparable to fine-tuning model weights while reducing computational costs

"prompt optimization emerging as a promising approach capable of delivering performance gains comparable to those achieved by fine-tuning model weights, while reducing computational costs in both optimization and serving"

Stronger teacher reasoning can partially substitute for optimizer-side search complexity in prompt optimization

"its advantage increases with stronger teacher models, suggesting that stronger teacher reasoning can partially substitute for optimizer-side search complexity"