DeepTRACE: Auditing Deep Research AI Systems for Tracking Reliability Across Citations and Evidence
Paper • 2509.04499 • Published
AI & ML interests
None defined yet.
Recent Activity
View all activity
PranavNVenkit
authored 10
papers 14 days ago
What if AI systems weren't chatbots?
Paper • 2605.07896 • Published • 8
The Need for a Socially-Grounded Persona Framework for User Simulation
Paper • 2601.07110 • Published
MMPersuade: A Dataset and Evaluation Framework for Multimodal Persuasion
Paper • 2510.22768 • Published • 8
LLMs Assist NLP Researchers: Critique Paper (Meta-)Reviewing
Paper • 2406.16253 • Published
Search Engines in an AI Era: The False Promise of Factual and Verifiable Source-Cited Responses
Paper • 2410.22349 • Published
Survey on Sociodemographic Bias in Natural Language Processing
Paper • 2306.08158 • Published
CALM : A Multi-task Benchmark for Comprehensive Assessment of Language Model Bias
Paper • 2308.12539 • Published
Training Reasoning Models to Generate Diverse Scientific Ideas
Paper • 2610.22324 • Published • 6
Best Friends, Not Forever: Evaluating Long-Horizon Persona Collapse and Behavioral Drift in AI Companions
Paper • 2607.28818 • Published • 1
shekkizh
authored a
paper 7 months ago
shekkizh
submitted a
paper to Daily Papers 7 months ago
shekkizh
authored a
paper 7 months ago
Post
2,504
🙋🏽♂️ Is your "multi agent" system really multi agentic? Or is it just a modular setup with a bunch of different prompts? 🤨
I’ve had this discussion way too often, so I finally wrote it all down. If you’re building with agents, you need to read this.
Here’s the TLDR:
✅ True multi agent systems require:
• Persistent, private state per agent
• Memory that impacts future decisions
• Adaptation based on past experiences
❌ Just having modular components, function calls, or multiple LLMs doesn't cut it. That’s not multi agentic. It’s just pipelining.
🤝 The magic is in evolving relationships, context retention, and behavioral shifts over time.
🧠 If your agents aren’t learning from each other or changing based on past experience… you are missing the point.
What do you think? Curious what patterns you're experimenting with 🧐
👉 Full post: https://shekkizh.github.io/posts/2025/04/multi-agents/
I’ve had this discussion way too often, so I finally wrote it all down. If you’re building with agents, you need to read this.
Here’s the TLDR:
✅ True multi agent systems require:
• Persistent, private state per agent
• Memory that impacts future decisions
• Adaptation based on past experiences
❌ Just having modular components, function calls, or multiple LLMs doesn't cut it. That’s not multi agentic. It’s just pipelining.
🤝 The magic is in evolving relationships, context retention, and behavioral shifts over time.
🧠 If your agents aren’t learning from each other or changing based on past experience… you are missing the point.
What do you think? Curious what patterns you're experimenting with 🧐
👉 Full post: https://shekkizh.github.io/posts/2025/04/multi-agents/
- 2 replies
Post
1,955
Think AGI is just around the corner? Not so fast.
When OpenAI released its Computer-Using Agent (CUA) API, I happened to be playing Wordle 🧩 and thought, why not see how the model handles it?
Spoiler: Wordle turned out to be a surprisingly effective benchmark.
So Romain Cosentino Ph.D. and I dug in and analyzed the results of several hundred runs.
🔑 Takeaways
1️⃣ Even the best computer-using models struggle with simple, context-dependent tasks.
2️⃣ Visual perception and reasoning remain major hurdles for multimodal agents.
3️⃣ Real-world use cases reveal significant gaps between hype and reality. Perception accuracy drops to near zero by the last turn 📉
🔗 Read our arxiv article for more details https://www.arxiv.org/abs/2504.15434
When OpenAI released its Computer-Using Agent (CUA) API, I happened to be playing Wordle 🧩 and thought, why not see how the model handles it?
Spoiler: Wordle turned out to be a surprisingly effective benchmark.
So Romain Cosentino Ph.D. and I dug in and analyzed the results of several hundred runs.
🔑 Takeaways
1️⃣ Even the best computer-using models struggle with simple, context-dependent tasks.
2️⃣ Visual perception and reasoning remain major hurdles for multimodal agents.
3️⃣ Real-world use cases reveal significant gaps between hype and reality. Perception accuracy drops to near zero by the last turn 📉
🔗 Read our arxiv article for more details https://www.arxiv.org/abs/2504.15434
- 3 replies
Post
624
Some interesting architectural choices made in Llama 4 models -- were these key to the 10M context? Possibly 🤔
🔍 Takeaways:
🧩 Interleaved Attention without position encoding
- LLaMA 4 removes explicit positional encoding in some attention layers to boost performance on longer contexts.
- The principles here could be similar to the residual connections to facilitate attention to early tokens without positional decay.
⚖️ Scaled Softmax to increase attention at inference time
- The max attention value (output of softmax) decreases as context size increases.
- Llama 4 incorporates a context-size dependent temperature in the softmax function to modify the slope of softmax, allowing the model to focus better on relevant tokens.
- Done only at inference time -- guessing it was more a choice after some observation on eval datasets.
What did you think of these choices?
🔍 Takeaways:
🧩 Interleaved Attention without position encoding
- LLaMA 4 removes explicit positional encoding in some attention layers to boost performance on longer contexts.
- The principles here could be similar to the residual connections to facilitate attention to early tokens without positional decay.
⚖️ Scaled Softmax to increase attention at inference time
- The max attention value (output of softmax) decreases as context size increases.
- Llama 4 incorporates a context-size dependent temperature in the softmax function to modify the slope of softmax, allowing the model to focus better on relevant tokens.
- Done only at inference time -- guessing it was more a choice after some observation on eval datasets.
What did you think of these choices?
jasonwu
authored a
paper over 2 years ago
Post
1,237
Hi folks,
Tenyx announced its latest model Llama3-TenyxChat-70B, which outperforms a GPT-4 variant on several MT-Bench measurements.
By post-training Llama-3 70B in 15 hours, our model improves reasoning capabilities leveraging the relationship between geometry and LLM task complexity (Take a look at our paper: https://arxiv.org/abs/2312.01648, to be presented at ICML 2024)
Model: tenyx/Llama3-TenyxChat-70B, HuggingFace Space: tenyx/Llama3-TenyxChat-70B
Tenyx announced its latest model Llama3-TenyxChat-70B, which outperforms a GPT-4 variant on several MT-Bench measurements.
By post-training Llama-3 70B in 15 hours, our model improves reasoning capabilities leveraging the relationship between geometry and LLM task complexity (Take a look at our paper: https://arxiv.org/abs/2312.01648, to be presented at ICML 2024)
Model: tenyx/Llama3-TenyxChat-70B, HuggingFace Space: tenyx/Llama3-TenyxChat-70B
jasonwu
authored a
paper over 3 years ago