Thinking LLMs: General Instruction Following with Thought Generation
Paper
β’
2410.10630
β’
Published
β’
18
Nexusflow
Enterprise
company
Verified
AI & ML interests
Democratize GenAI Agents for Enterprise Workflows
Recent Activity
View all activity
Nexusflow's activity
nexus-jt-llmΒ
authored
a
paper
3 months ago
banghuaΒ
authored
a
paper
6 months ago
banghuaΒ
authored
a
paper
8 months ago
Post
1528
Have we really squeezed out the capacity of a compact chat model? Thrilled to see our latest open model, Starling-7B, ranks 13th among all models in Chatbot Arena!
π As a 7B model, Starling surpasses larger open and proprietary models, including Claude-2, GPT-3.5-Turbo, Gemini Pro, Mixtral 8x7B and Llama2-70B, and is currently the best 7B chat model in Chatbot Arena!
Try out the model on HF here: Nexusflow/Starling-LM-7B-beta
π As a 7B model, Starling surpasses larger open and proprietary models, including Claude-2, GPT-3.5-Turbo, Gemini Pro, Mixtral 8x7B and Llama2-70B, and is currently the best 7B chat model in Chatbot Arena!
Try out the model on HF here: Nexusflow/Starling-LM-7B-beta
banghuaΒ
authored
a
paper
10 months ago
Post
π Exciting breakthrough in LLM reliability! π§ NexusRaven-V2, our cutting-edge function-calling LLM, has set a new standard in minimizing AI hallucinations, surpassing GPT-4's performance in a recent third-party independent research benchmark.
Dive into our latest blog post to explore how we're pioneering reliable agents with minimal hallucinations: https://nexusflow.ai/blogs/towards-reliable-agents-with-minimal-hallucination
Key Highlights:
π Zero Hallucinations: NexusRaven-V2 showcased remarkable accuracy with zero hallucinations in 840 tests, focusing on tool selection and usage β a significant leap over GPT-4 with 23 hallucinations.
π Enhanced Success Rates: It boasts a 9% higher success rate than GPT-4 in information-seeking applications requiring meticulous attention to detail and a 4% increase in adversarial scenarios that demand a deep understanding of tool documentation, even with vague tool and API argument names.
Try NexusRaven-V2 on Huggingface: Nexusflow/NexusRaven-V2-13B
Check out the original third-party benchmark: https://arxiv.org/abs/2401.08326
Dive into our latest blog post to explore how we're pioneering reliable agents with minimal hallucinations: https://nexusflow.ai/blogs/towards-reliable-agents-with-minimal-hallucination
Key Highlights:
π Zero Hallucinations: NexusRaven-V2 showcased remarkable accuracy with zero hallucinations in 840 tests, focusing on tool selection and usage β a significant leap over GPT-4 with 23 hallucinations.
π Enhanced Success Rates: It boasts a 9% higher success rate than GPT-4 in information-seeking applications requiring meticulous attention to detail and a 4% increase in adversarial scenarios that demand a deep understanding of tool documentation, even with vague tool and API argument names.
Try NexusRaven-V2 on Huggingface: Nexusflow/NexusRaven-V2-13B
Check out the original third-party benchmark: https://arxiv.org/abs/2401.08326
banghuaΒ
authored
a
paper
about 1 year ago
banghuaΒ
authored
7
papers
over 1 year ago
Bridging Offline Reinforcement Learning and Imitation Learning: A Tale of Pessimism
Paper
β’
2103.12021
β’
Published
Doubly Robust Self-Training
Paper
β’
2306.00265
β’
Published
β’
1
On Optimal Caching and Model Multiplexing for Large Model Inference
Paper
β’
2306.02003
β’
Published
β’
1
Fine-Tuning Language Models with Advantage-Induced Policy Alignment
Paper
β’
2306.02231
β’
Published
β’
2
Principled Reinforcement Learning with Human Feedback from Pairwise or $K$-wise Comparisons
Paper
β’
2301.11270
β’
Published
β’
2
Online Learning in Stackelberg Games with an Omniscient Follower
Paper
β’
2301.11518
β’
Published
β’
1
Jump-Start Reinforcement Learning
Paper
β’
2204.02372
β’
Published
β’
1