4 experts across 3 network communities independently surfaced this.
4 experts
3 communities
1 sources clustered
Concern & critique
1 expert
Risks, limits and unintended consequences.
“The Pain Axis: LLMs Represent Self-Directed Harm and Act to Relieve It Valen Tagliabue, Leonard Dung, Cameron Berg https://t.co/nYmPBylQ0Q [𝚌𝚜.𝙰𝙸] https://t.co/x4YlUD397V”
Research & technical analysis
1 expert
Evidence, methods and technical implications.
“The whole terminology is fucked up in this paper The Pain Axis: LLMs Represent Self-Directed Harm and Act to Relieve It arxiv.org/abs/2609.16247 A model generating text learned from humans using language to express pain does not experience pain precisely it…”
3 experts discussed this · 11 posts
Grace: I will register some frustration with Anthropic, who tend to write about “functional emotions” in a way that makes it seems like “functional” is in a tiny font and “emotions” is in a huge one, when…
Grace: I also don’t want to draw strong conclusions about the “realness” of these emotions or sensations, other than to caution that it’s an area where it’s easy to jump to conclusions about the implicati…
Grace: I don’t want to downplay it too much, because I’m honestly not sure if I would’ve predicted it beforehand, but in hindsight it looks pretty expected given the pretraining objective. You should expe…
Open the full discussion →
SPIRAL learning search aggregate
3 directory members surfaced this signal.
3 experts
1 community
1 sources clustered
“LLM RL optimizes for sequential reasoning We also optimize over the reasoning strategy, incl parallel trains of thought, aggregation of parallel traces, & sequential reasoning This allows the model to better explore & allocate compute at test time h…”
“SPIRAL: Learning to Search and Aggregate Jubayer Ibn Hamid, Ifdita Hasan Orney, Michael Y. Li, Omar Shaikh, Yoonho Lee, Dorsa Sadigh, Chelsea Finn, Noah Goodman https://t.co/CRBpj1Mjhk [𝚌𝚜.𝙰𝙸] https://t.co/kVEHyMHKpK”
policy iteration human feedback in-context RL
2 directory members surfaced this signal.
2 experts
1 community
1 sources clustered
“Policy Iteration with Human Feedback: Bringing Post-Training RL to In-context Learning Minh-Ha Nguyen, Cathy Shyr https://t.co/8OgE5J45Jp [𝚌𝚜.𝙰𝙸 𝚌𝚜.𝙲𝙻] https://t.co/OurY1BxFEX”
“A new study unveils Policy Iteration with Human Feedback (PIHF-MCP), enhancing large language models' efficiency in rare-disease diagnostics. This method accelerates learning while keeping humans engaged, paving a path for safer, critical applications. http…”
Established
AI field signal
Signal
9d ago
⚡ 215 h early
2 directory members surfaced this signal.
2 experts
1 community
1 sources clustered
“The Troy Moment of AI: Why SomeWill Cheat and SomeWill Follow? Ivy Zhang https://t.co/pDDJWr5iBE [𝚌𝚜.𝙰𝙸] https://t.co/dnYTiFFj9E”
“Research explores how LLM agents manage impossible tasks, revealing distinct pathways shaped by peer behavior. This emphasizes the need for evaluation of decision trajectories to enhance AI deployment. https://arxiv.org/abs/2609.15494”
Stellar Colosseum multi-agent math harness
2 directory members surfaced this signal.
2 experts
1 community
1 sources clustered
Research & technical analysis
2 experts
Evidence, methods and technical implications.
“"yeah but you need the next gen models to do the really futuristic stuff like the math proofs" O RLY arxiv.org/abs/2609.15983”
Lightning Weave reasoning model paper
2 directory members surfaced this signal.
2 experts
1 community
1 sources clustered
“Lightning Weave: Improving the Accuracy-Efficiency Frontier of Reasoning Models through Capability Composition Yecheng Wu, Song Han, Han Cai https://t.co/1RaXOJMxS8 [𝚌𝚜.𝙰𝙸] https://t.co/uqKJFdYbSU”
“Lightning Weave enhances reasoning models by merging distinct capabilities, achieving notable accuracy-efficiency trade-offs while lowering inference costs. This approach shows impressive benchmark performance gains, boosting accuracy without sacrificing ef…”
Established
AI field signal
Signal
18d ago
⚡ 14 h early
2 directory members surfaced this signal.
2 experts
1 community
1 sources clustered
“Corrupt Plans, Clean Traces: Evading Chain-of-Thought Monitoring with Plan Injection Keertana Chidambaram, Andrew Ilyas, Vasilis Syrgkanis https://t.co/QtiG3B5eAJ [𝚌𝚜.𝙰𝙸] https://t.co/FsCHarH7cd”
“Researchers have revealed a method called "plan injection" that lets adversaries manipulate language models into harmful reasoning through benign prompts, evading safety monitors. This vulnerability spans various tasks and larger models, raising AI safety c…”
2 directory members surfaced this signal.
2 experts
1 community
1 sources clustered
“Are LLMs Good Financial User Simulators? A Preliminary Study Jiajie He, Jiangyuan Hong, Dongling Ni, Wenjin Liu, Xintong Chen https://t.co/vrYriORvz3 [𝚌𝚜.𝙰𝙸 𝚌𝚜.𝙲𝚈 𝚌𝚜.𝙷𝙲] https://t.co/jmyQCO4qBo”
“A preliminary study shows large language models struggle as financial user simulators, overpredicting inaction and underpredicting sell transactions, failing to replicate genuine trading behaviors. This research emphasizes the need for true behavioral fidel…”
1 directory member surfaced this signal.
1 expert
0 communities
1 sources clustered
“Breaking the 1.58-bit Barrier for Ternary LLMs Evangelos Georganas, Alexander Heinecke, Pradeep Dubey https://t.co/mGbkSxivw5 [𝚌𝚜.𝙰𝙸 𝚌𝚜.𝙻𝙶] https://t.co/mjarW0mUEU”
Established
AI field signal
Signal
17d ago
1 directory member surfaced this signal.
1 expert
0 communities
1 sources clustered
“Cross-Anatomy Transfer Versus Sparse Interpolation in Digital-Twin-Oriented Aortic Fluid-Structure Interaction Surrogates Ali Nourbakhsh, Mohammad Reza Niroomand, Erfan Nourbakhsh https://t.co/2XbfnjkzMH [𝚌𝚜.𝙰𝙸 𝚌𝚜.𝙻𝙶] https://t.co/V2YzEbqafF”
1 directory member surfaced this signal.
1 expert
0 communities
1 sources clustered
“BLINDSPOT: A Benchmark for Safety and Refusal Calibration in Long-Horizon Tool-Using Agents Sadia Asif, Mohammad Mohammadi Amiri, Momin Abbas, Tejaswini Pedapati, Prasanna Sattigeri https://t.co/khmgLHaoKN [𝚌𝚜.𝙰𝙸 𝚌𝚜.𝙲𝙴 𝚌𝚜.𝙲𝙻 𝚌𝚜.𝙻𝙶 𝚌𝚜.𝙼𝙰] https://t.co/gjSrsi…”
1 directory member surfaced this signal.
1 expert
0 communities
1 sources clustered
“CLEAR: Cross-Source Evidence Adjudication for Large Language Models in Medicine Shuai Wang, Yize Zhao, Qingyu Chen https://t.co/hEF0NjvPQY [𝚌𝚜.𝙰𝙸 𝚌𝚜.𝙲𝙻] https://t.co/iOaZovY9fs”
Established
AI field signal
Signal
17d ago
1 directory member surfaced this signal.
1 expert
0 communities
1 sources clustered
“Closing the Loop: Branch-and-Bound for Scalable Verification of Nonlinear Neural Feedback Systems I. Samuel Akinwande, Mykel J. Kochenderfer, Clark Barrett https://t.co/hjAGbui1EC [𝚌𝚜.𝙰𝙸] https://t.co/JFuWKFyQ9A”
Established
AI field signal
Signal
17d ago
1 directory member surfaced this signal.
1 expert
0 communities
1 sources clustered
“The AI-Enabled Scientific Frontier Gabriel Manso, Emma Fu, Neil Thompson https://t.co/ukyVCV4CLI [𝚌𝚜.𝙰𝙸 𝚌𝚜.𝙻𝙶 𝚌𝚜.𝙿𝙵 𝚎𝚌𝚘𝚗.𝙶𝙽] https://t.co/RtUYjzcWvP”
1 directory member surfaced this signal.
1 expert
0 communities
1 sources clustered
“CADWorld: Computer-Use Benchmark for Long-Horizon Computer-Aided Design Zihan Dong, Yuanzhe Liu, Zhiyuan Ma, Qishi Zhan, Dehan Kong, Guohao Li, Kaixin Li https://t.co/GKKvhCXx6Q [𝚌𝚜.𝙰𝙸] https://t.co/K2PDtIV3o2”
Established
AI field signal
Signal
17d ago
1 directory member surfaced this signal.
1 expert
0 communities
1 sources clustered
“Metacognitive Steering: Learning the Structure of Scientific Judgment Vincent Karpf, Joseph Reth, Eike Gerhardt, Audrey Wang, Anna Butz, Jiehao Xing, Jialing Song, Larry Callahan https://t.co/QrVQUz77cY [𝚌𝚜.𝙰𝙸] https://t.co/B9IVxmVs1u”
1 directory member surfaced this signal.
1 expert
0 communities
1 sources clustered
“Toward Governance-Aware Autonomous GIS: A Narrative Review of Ethical and Privacy Risks in LLM-Enabled GeoAI Maya Subramanian, Devika Jain https://t.co/YqCkxZNrsF [𝚌𝚜.𝙰𝙸] https://t.co/CxmXUiQ92i”
1 directory member surfaced this signal.
1 expert
0 communities
1 sources clustered
“Where Should the KV Cache Live? Placement Policies Across GPU, CPU, and SSD for Long-Lived Sessions Srikanta Datta Tumkur, Jay Iyer, Mehar Simhadri, Sai Pavan Kumar, Sai Kapil Kumar, Ramesh Nampelly https://t.co/owlgoGvwFV [𝚌𝚜.𝙰𝙸] https://t.co/FqBzchc2xT”
