rong wang

@wangrongbily

coder

Joined March 2022

8Posts 0Followers 41Following

rong wang reposted

LangChain

@LangChainAI

Feb 18

🐘 LangMem SDK: long-term memory for AI agents Different types of memory enable different agent capabilities. We're releasing an open-source library that implements some of these patterns so you can build agents that learn from interactions. The LangMem SDK makes it easier to:…

LangChainAI's tweet image. 🐘 LangMem SDK: long-term memory for AI agents

Different types of memory enable different agent capabilities. We're releasing an open-source library that implements some of these patterns so you can build agents that learn from interactions. The LangMem SDK makes it easier to:…

rong wang reposted

echo.hive

@hive_echo

Jan 19

Memory systems for LLMs are easy 🥳 here are 5 implementations as objects 1) Vector memory 2) summary memory 3) time window memory 4) keyword memory 5) hierarchical memory 1)Vector Memory System ------------------- This memory system uses OpenAI's embeddings to create vector…

hive_echo's tweet image. Memory systems for LLMs are easy 🥳
here are 5 implementations as objects

1) Vector memory
2) summary memory
3) time window memory
4) keyword memory
5) hierarchical memory

1)Vector Memory System
-------------------
This memory system uses OpenAI's embeddings to create vector…

rong wang reposted

Unwind AI

@unwind_ai_

Feb 19

AI Agents get long-term memory with knowledge graphs. Graphiti lets you build AI Agents that learn from past interactions and evolve over time. Works with unstructured text and JSON. 100% opensource.

rong wang reposted

Kaixuan Huang

@KaixuanHuang1

Feb 11

Do LLMs have true generalizable mathematical reasoning capability or are they merely memorizing problem-solving skills? 🤨 We present MATH-Perturb, modified level-5 problems from MATH dataset to benchmark LLMs' generalizability to slightly perturbed problems. 🔗…

KaixuanHuang1's tweet image. Do LLMs have true generalizable mathematical reasoning capability or are they merely memorizing problem-solving skills? 🤨

We present MATH-Perturb, modified level-5 problems from MATH dataset to benchmark LLMs' generalizability to slightly perturbed problems.

🔗…

rong wang reposted

Rohan Paul

@rohanpaul_ai

Jan 1

WorkflowLLM enables LLMs to handle 70+ action workflows, a 10x improvement over current capabilities An LLM that can orchestrate real-world automation workflows at production scale Original Problem 🤔: Current LLMs can only handle small workflows with around 6 actions and…

rohanpaul_ai's tweet image. WorkflowLLM enables LLMs to handle 70+ action workflows, a 10x improvement over current capabilities

An LLM that can orchestrate real-world automation workflows at production scale

Original Problem 🤔:

Current LLMs can only handle small workflows with around 6 actions and…

rong wang reposted

amit siwal

@amitsiwal

Feb 20

Hybrid Approach: LLM controls the execution but by expressing "entire plan" as js code which is then executed using an AST runtime

amitsiwal's tweet image. Hybrid Approach: LLM controls the execution but by expressing "entire plan" as js code which is then executed using an AST runtime

rong wang reposted

Tom Dörr

@tom_doerr

Feb 5

LLM framework for task decomposition and agents

rong wang reposted

Jason Weston

@jaseweston

Jan 31

💭🔎 Introducing EvalPlanner – a method to train a Thinking-LLM-as-a-Judge that learns to generate planning & reasoning CoTs for evaluation. Strong performance on RewardBench, RM-Bench, JudgeBench & FollowBenchEval. Paper 📄: arxiv.org/abs/2501.18099