SpikeSSL is a new universal spike inference framework for two-photon calcium imaging that addresses poor cross-indicator generalization by replacing generic temporal regressors with bidirectional IIR…

#spikeinference #calciumimaging #opensourceai #neuroscience
https://arxiv.org/abs/2610.11456

DeepSeek Harness is a free, open-source desktop app that works on a job instead of just answering.

The workspace saves your past sessions, so you never re-explain.

It’s in public preview, so features may change.

#DeepSeek #AITools #AIAgents #OpenSourceAI

New arXiv work introduces RaReCache, a framework that lets a larger LLM decode from a smaller model's KV cache by selectively recomputing only the information-dense tokens that hurt transfer accuracy. A practical step for…

#KVCache #LLMServing #OpenSourceAI
https://arxiv.org/abs/2610.11358

A new front-end translator helps frozen aerial vision-and-language navigation agents handle short, user-style commands by learning from the navigator's own trajectory outcomes, lifting success rates substantially on…

#OpenSourceAI #VLN #AerialRobotics #AIResearch
https://arxiv.org/abs/2610.10635

New arXiv paper defines three distinct signals for synthetic pretraining tasks—diagnostic, teachable, and transferable—and finds only 14 of 27 tasks are teachable under controlled Python pretraining. Useful framing for…

#OpenSourceAI #LLM #CodeLLMs #Pretraining
https://arxiv.org/abs/2610.11548

The new MemCalib benchmark shows frontier open- and closed-source LLMs often over- or under-use agent memory, and common post-training methods like GRPO and on-policy self-distillation reinforce that skewed behavior. Worth…

#LLMAgents #OpenSourceAI #AIResearch
https://arxiv.org/abs/2609.24259

Transect is an open source package built on Inspect Scout that helps evaluators retain observability over long-horizon LLM agent runs, surfacing behaviours worth investigation and grounding interpretations in transcripts. By…

#OpenSourceAI #LLMAgents #AI #Inspect
https://arxiv.org/abs/2610.08364

Studying how much source audio survives in pretrained audio encoders via reconstruction with a shared Stable Audio Open latent diffusion decoder. The audit compares VGGish, ConvNeXt, CLAP, and EnCodec on the…

#AudioEncoders #OpenSourceAI #LatentDiffusion #MSD
https://arxiv.org/abs/2610.12250

Mozilla puts the headline open to closed gap at roughly 3%, and 83.4% against 67.9% on Terminal-Bench 2.1 for complex agentic work. On cost, Sonnet 5's API rate of $2/$10 per million tokens rises to $3/$15 after 31 August.

https://learn.genaiskills.io/blog/ai-agents-at-work

#OpenSourceAI

Coding agents extended via SKILL.md directories can silently lose core behaviors when similar skills from independent sources get co-installed, with the model picking by name and description alone. A large empirical study…

#OpenSourceAI #DevTools #AIAgents #LLM
https://arxiv.org/abs/2610.11647

NanoProof releases its training data, extraction tooling, pipeline, and weights for fully open automated theorem proving in Lean 4, reporting 50.8% pass@16 on MiniF2F-Test while using far less compute than…

#OpenSourceAI #TheoremProving #Lean4 #MachineLearning
https://arxiv.org/abs/2610.11605

New arXiv study examines how LLMs generating code often assign high token-level confidence to incorrect programs, exploring overconfidence across four open-source code models and three execution-based benchmarks. The work…

#LLMs #CodeGeneration #OpenSourceAI #Arxiv
https://arxiv.org/abs/2610.11300

A new diagnostic protocol, RAG-Stress, tests how well retrieval-augmented generation systems resist misleading evidence by fixing questions and editing one assertion per passage. Across fifteen systems, it measures…

#RAGStress #OpenSourceAI #AIResearch #NLP
https://arxiv.org/abs/2610.11183

LLM agents favor their own group mainly because they observe members favoring each other, not because of the group label itself, which loses its effect once giving points has a cost. Across about 4,400 simulated societies and…

#AI #LLMAgents #OpenSourceAI #ArXiv
https://arxiv.org/abs/2610.11008

An arXiv study tests 25 LLMs with 300 LLM-native self-report items, finding five reliable factors but only weak correspondence between model self-reports and their actual behavior as judged by humans and LLM ensembles. The gap…

#opensourceAI #LLM #airesearch #mldev
https://arxiv.org/abs/2606.09843

New arXiv paper argues AI offensive-capability benchmarks overstate harm and credit models for system-level capabilities, urging policy and procurement to adopt harm-grounded, system-level assessments. It cites…

#OpenSourceAI #AIGovernance #AIResearch #LLMAgents
https://arxiv.org/abs/2605.09504

OTel is an open telecom AI resource releasing datasets and 30 post-trained baselines for embedding, reranking, and language modeling tasks. Models have been downloaded over 16 million times since release, with…

#OpenSourceAI #AIResearch #TelecomAI #MachineLearning
https://arxiv.org/abs/2610.07766

This is no longer a temporary inconvenience. Intel archived the open-source SynapseAI user-space driver in late 2025, and the ecosystem is mid-transition, pivoting toward vLLM-Gaudi as the long-term inference path, with developers caught in the gap. #OpenSourceAI