SACHIN KUMARPairDistill: Pairwise Relevance Distillation to enrich training of dense retrieval modelsTo further enhance retrieval performance, knowledge distillation techniques primarily distill knowledge from pointwise rerankers which…20h ago20h ago
SACHIN KUMARGemFilter : using earlier LLM layers as filters for faster Long-Context LLMs with 1000x Input…LLMs capabilities in handling long context inputs, comes at the cost of increased computational resources and latency. To address it, this…3d ago3d ago
SACHIN KUMARRULER approach: improving LLM instruction following to generate responses of a specified lengthWhen required to generate responses of a specific length, large language models often struggle to meet users’ needs due to their inherent…3d ago3d ago
SACHIN KUMARBEATS framework : improving LLMs Mathematical capabilities with back-verification and Pruning Tree…Even with usage of techniques such as supervised fine-tuning (SFT), prompt engineering, and search-based methods to improve the…4d ago4d ago
SACHIN KUMARMinstrel: tool for structural prompt generation with Multi-Agents Coordination based on LangGPT…In this paper[1], authors introduce Minstrel, a multi-generative agent system with reflection to automate the generation of structural…5d ago5d ago
SACHIN KUMARLogic-of-Thought : prompting approach leveraging propositional logic to enhance logical reasoning…Prompting techniques such as Chain-of-Thought, can improve the reasoning ability of LLMs to some extent, they suffer from an unfaithful…6d ago6d ago
SACHIN KUMARIteration of Thought: multi-agent framework to enhance LLM Reasoning capabilitiesUsing well-structured prompts in a conversational manner, human users can effectively influence an LLM to develop more thoughtful and…Sep 20Sep 20
SACHIN KUMARMagicore: Multi Agent Iteration Framework for Coarse-to-fine Refinement for improved LLM solution…Large Language Models’ (LLM) reasoning can be improved using test-time aggregation strategies, i.e., generating multiple samples for each…Sep 20Sep 20
SACHIN KUMARPromptriever: first Zero-Shot Promptable Instruction-Trained Retriever ModelIn this paper[1], authors present Promptriever, the first retrieval model able to be prompted with natural language prompts like an…Sep 19Sep 19
SACHIN KUMARDiagram of Thought (DoT) : framework that models iterative reasoning in LLMs as construction of a…Traditional approaches that represent reasoning as linear chains or trees, however, DoT organizes propositions, critiques, refinements…Sep 171Sep 171
