6 citations · 9 across the 10 of their papers we have counts for
12 papers
Optimizing CUDA like a Human: Micro-Profiling Tools as Expert Surrogates for LLM-Based GPU Kernel Optimization
Jiading Gai, Shuai Zhang, Kaj Bostrom +6
We present KernelPro, a closed-loop multi-agent system that automatically generates, profiles, and iteratively optimizes GPU kernel code by integrating large language model (LLM) c…
ReSyn: Autonomously Scaling Synthetic Environments for Reasoning Models
Andre He, Nathaniel Weir, Kaj Bostrom +4
Reinforcement learning with verifiable rewards (RLVR) has emerged as a promising approach for training reasoning language models (RLMs) by leveraging supervision from verifiers. Al…
Generating Data-Driven Reasoning Rubrics for Domain-Adaptive Reward Modeling
Kate Sanders, Nathaniel Weir, Sapana Chaudhary +2
An impediment to using Large Language Models (LLMs) for reasoning output verification is that LLMs struggle to reliably identify errors in thinking traces, particularly in long out…
MaxCode: A Max-Reward Reinforcement Learning Framework for Automated Code Optimization
Jiefu Ou, Sapana Chaudhary, Kaj Bostrom +4
Large Language Models (LLMs) demonstrate strong capabilities in general coding tasks but encounter two key challenges when optimizing code: (i) the complexity of writing optimized…
VeriCoT: Neuro-symbolic Chain-of-Thought Validation via Logical Consistency Checks
Yu Feng, Nathaniel Weir, Kaj Bostrom +5
LLMs can perform multi-step reasoning through Chain-of-Thought (CoT), but they cannot reliably verify their own logic. Even when they reach correct answers, the underlying reasonin…
Lil-Bevo: Explorations of Strategies for Training Language Models in More Humanlike Ways
Venkata S Govindarajan, Juan Diego Rodriguez, Kaj Bostrom +1
We present Lil-Bevo, our submission to the BabyLM Challenge. We pretrained our masked language models with three ingredients: an initial pretraining with music data, training on sh…
MuSR: Testing the Limits of Chain-of-thought with Multistep Soft Reasoning
Zayne Sprague, Xi Ye, Kaj Bostrom +2
While large language models (LLMs) equipped with techniques like chain-of-thought prompting have demonstrated impressive capabilities, they still fall short in their ability to rea…
Deductive Additivity for Planning of Natural Language Proofs
Zayne Sprague, Kaj Bostrom, Swarat Chaudhuri +1
Current natural language systems designed for multi-step claim validation typically operate in two phases: retrieve a set of relevant premise statements using heuristics (planning)…
Natural Language Deduction with Incomplete Information
Zayne Sprague, Kaj Bostrom, Swarat Chaudhuri +1
A growing body of work studies how to answer a question or verify a claim by generating a natural language "proof": a chain of deductive inferences yielding the answer based on a s…
Natural Language Deduction through Search over Statement Compositions
Kaj Bostrom, Zayne Sprague, Swarat Chaudhuri +1
In settings from fact-checking to question answering, we frequently want to know whether a collection of evidence (premises) entails a hypothesis. Existing methods primarily focus…
Flexible Generation of Natural Language Deductions
Kaj Bostrom, Xinyu Zhao, Swarat Chaudhuri +1
An interpretable system for open-domain reasoning needs to express its reasoning process in a transparent form. Natural language is an attractive representation for this purpose --…
Byte Pair Encoding is Suboptimal for Language Model Pretraining
Kaj Bostrom, Greg Durrett
The success of pretrained transformer language models (LMs) in natural language processing has led to a wide range of pretraining setups. In particular, these models employ a varie…