BotPsy pairs an on-device assistant with a small agentic model
The Hugging Face post describes a personal assistant built around a language model under one billion parameters.
Stories per day · 30 days
2 RSS
Relatedon-device4open-weights4agents2hugging-face2huggingface2spark2acoustic-attack1botpsy1byte-level1carbon-x1clinical-decision-networks1diffusion-llm1dna1fern1genomics1gguf1inception1keyboard1
21 stories, newest first · one card per story
Stories tagged #language-model
Paper says domain-invariant models recover keystrokes from keyboard audio
The authors say inference still works on unseen keyboards, users, noise, and compressed channels when a language model corrects predictions.
Meta launches Muse Ember, a small model for agent routine tasks
Meta prices it at $0.10 per million input tokens and says it scores 36 on Artificial Analysis for about $0.14 a task.
Language model audits and repairs flaws in small clinical decision networks
Claude Opus 5.5 audited and repaired flaws in radiotherapy networks with 433-4801 parameters
DeepMem advances byte-level language modeling with multiscale architecture
DeepMem separates semantic compression from exact byte preservation using a learned invertible subband transform to create a hierarchy of shorter sequences.
Dory1.0-4B looped language model released
4-billion-parameter looped model outperforms larger models on benchmarks with open checkpoints.
Mirendil launches Fern 0.1, its first model, with 10M-token context
The captured page is only a cookie notice, so architecture, weights, and access are not described.
Hugging Face releases Carbon-X, a 14B open DNA language model
The announcement frames Carbon-X as an open model that jointly handles DNA sequences and natural language.
1
Inception releases Mercury 3 diffusion LLM at 1,068 tokens per second
The page body is a cookie banner, so only the stated generation speed is available, not architecture or availability.
1
XHToken releases Spark-X2.6-4B GGUF for on-device use
The 4B model keeps the Spark-X2.5 architecture and claims coding and agent gains from post-training alone.
XHToken publishes Spark-X2.6-4B on-device model weights
The 4B model keeps the Spark-X2.5 architecture and adds none, low, and high thinking levels.
Muse Glimmer 30B released as an on-device agentic model
The Apache 2.0 model takes images, runs at about 4-bit on one GPU, and claims to beat same-size peers.
Language Models that Play Chess and Explain Their Moves
Queen, a 4B-parameter chess-language model, reaches 2697 Elo via iterative distillation while explaining moves, beating frontier models with ~1000x fewer parameters.
Open-sourcing AstaBrief, the fast report-generation model in Asta
Ai2 open-sourced AstaBrief, the fast report-generation model used in Asta.
Language Models that Play Chess and Explain Their Moves
Queen, a 4B chess language model, reaches Grandmaster strength while explaining its moves and plans.
PSSA: A non-transformer language model written from scratch in Rust
PSSA, a non-transformer language model written from scratch in Rust, drew modest Hacker News discussion.
FactorEngram: Factorized N-gram Memory with Basis-Level Gating for Language Models
FactorEngram adds factorized n-gram memory with basis-level gating to language models.
Requirement-Bound Verified Commissioning: A Frozen Four-Billion-Parameter Local Model as a Candidate Generator under an External Acceptance Layer with Verification and Release Authority
A frozen four-billion-parameter model proposes commissioning plans, but a sealed external gate decides every release.
Can gzip be a language model?
A developer shows gzip can continue text by beam-searching byte sequences that compress best against a primed corpus.
Pinocchio: Fast Uncertainty Estimates for Black-Box Language Models
Pinocchio introduces an external calibrator for black-box LLM API models with 0.862 AUROC and zero-shot transfer to thirteen unseen models.
Due to concerns about malicious applications, GPT2 will not be released (2019)
OpenAI's landmark 2019 GPT-2 post withheld the full 1.5B-parameter model over misuse concerns, releasing only a smaller variant and paper.
5