Reading room.
Engineering blogs & AI research
A reading room from engineering teams and AI labs: titles, short feed excerpts, and links to the originals. For my own essays, explore the writing archive.
- 1359 articles tracked
- 37 sources
- Refresh scheduled daily
Sources
Continual-ARC: Measuring Skill Acquisition over a Working Lifetime
Agent harnesses have emerged as a promising mechanism for continual learning, alongside test-time parametric learning. However, the current benchmarks have several limitations: they generally focus on a single applied do…
When Does the Real Wiring Win? Testing connectome structure-function claims against unmeasured physiology
Connectome-constrained models are often used to argue that a measured wiring diagram outperforms shuffled or rewired wiring, and hence that structure shapes function. Such comparisons can confound wiring with how the mod…
Pyrite: Building and Benchmarking Protein-Ligand Pose Ranking Agents
Predicting how a ligand binds to its target is central to small-molecule drug discovery. Modern cofolding models can generate diverse candidate poses, but reliably ranking these candidates remains challenging. We introdu…
Contrastive Discovery: Open-Ended Scientific Discovery over Competing Explanations
LLM agents are increasingly used for open-ended, data-driven scientific discovery, yet most of them evaluate hypotheses pointwise, testing an empirical claim $h$ only against its complement $\neg h$. When evidence reject…
AutoScaffold: AI-Guided Scaffolding for Reinforcement Learning of Language Models
Reinforcement learning with verifiable rewards (RLVR) produces rich training experience, yet outcome rewards reduce each attempt to a single number and leave most of this information unused. Training-time scaffolds, such…
TorchCurves: Differentiable Parametric Curves
We present TorchCurves, a PyTorch library that treats vector-valued parametric curves as learnable primitives. Its central operation maps batches of scalar inputs to points on curves represented in B-spline or Legendre p…
Synapse-Control: Making Agent Experience Actionable with Lightweight Runtime Control
Large language model agents reuse execution experience to improve reliability on complex tool-based tasks. Systems such as Sentry call large models to generate failure diagnoses before retrieving advice, adding inference…
An Interpretable Stylometric Framework for Early Chinese Buddhist Translator Styles: A Case Study on the Attribution of Disputed Works of An Shigao
Attributing early Chinese Buddhist translations requires evidence that scholars can trace back to the texts. We develop an interpretable stylometric framework that uses features defined or reviewed by Buddhist philologis…
OpticalRAG: Pixel-Space Compression for Token-Efficient Retrieval-Augmented Generation
Retrieval-augmented generation (RAG) is bottlenecked by the token budget of large language models: feeding long documents into the context window is expensive, while aggressive textual compression discards fine-grained i…
You Are What You Read: Indian Political News Media Bias Dataset for Hindi Language
Media bias understanding and detection remains an under-researched problem for Indian context, specifically for Indic languages, such as, Hindi, which is widely spoken in India. In this paper, we introduce a comprehensiv…
Walrus Technical Report
We present Walrus, a unified 8B speech-and-audio foundation model handling speech recognition (including code-switching and multi-talker speech), speech-to-text and speech-to-speech translation, text-to-speech, emotion r…
Do Language Models Feel the Sentiment? Investigating Emotion Vectors
Emotion vectors are directions in a language model's internal representations that are associated with emotional concepts. Prior work shows that these vectors can influence model behavior, but their role in sentiment cla…
Word-Level Alignment of OCR and Ground Truth under Zoning and Reading-Order Errors
OCR post-processing needs OCR text aligned word by word with its ground truth. Data sets rarely provide that mapping, and a monotone aligner can lose its place when zoning errors permute the reading order. We revisit thi…
LLM Validation Makes Minimal-Pair Benchmarks Easier for Language Models
Minimal-pair benchmarks are only as good as their labels, and LLMs are an obvious substitute for costly human validation. We test it on BLiMP (English) and QFrBLiMP (Quebec French), the two public minimal-pair datasets r…
Antalia-2 Mini: A 7.6-million-parameter Turkish text-to-speech model trained entirely on synthetic speech
Antalia-2 Mini is a single-voice Turkish text-to-speech model with 7,620,931 parameters. It reads characters and produces 48 kHz audio at 41 times real time on an Apple M5 laptop CPU (2 threads). Its training audio came…
PolyRoute: A Twelve-Language and Code-Mixed Benchmark for Typed-Decision Routing, with Baselines and a Leaderboard
Typed-decision encoders answer a question about a text in one forward pass, but their behaviour outside English is rarely measured on a fixed yardstick. We release PolyRoute: 322 native-language support messages in 12 la…
Alania-2: Turkish text-to-speech with instant and fine-tuned voice cloning
We present Alania-2, a Turkish text-to-speech (TTS) system with two built-in voices, voice design from text descriptions, and two kinds of voice cloning. Alania-2 is a 488M-parameter model with the VoxCPM architecture, w…
Beyond Passive Imitation: On-Policy Mid-Training Distillation for Language Models
Continual pre-training (CPT) adapts large language models (LLMs) to specialized domains, yet its standard next-token prediction (NTP) objective treats every corpus token as a fixed target, regardless of its utility to th…
Are Causal Neuron Circuits Unique and Transferable?
Interpretability demonstrations often show that intervening on a group of neurons changes a model's answer. A recent example removes an 18-neuron group in Qwen3-32B and the model answers "Simba" instead of "Scar" to a tw…
Beyond Fluent Responses: Finding User Dissatisfaction in Human-AI Interactions
As large language models generate increasingly high-quality responses, recent work on response evaluation has moved beyond response-centered criteria, such as correctness and coherence, toward user-centered criteria, inc…
AVCC: Audio-Visual Conversation Corpus for Turn-Taking Prediction with a Human Benchmark
} Predicting who will speak next, and when, is a prerequisite for conversational agents and social robots that take turns without waiting for silence. Multiparty next-speaker prediction (NSP) has been studied on corpora…
Understanding Mixture-of-Experts Language Models Beyond Expert Boundaries
In Mixture-of-Experts (MoE) models, the nature of expert specialization remains controversial. Reports of domain-associated expert preferences contrast with analyses attributing routing primarily to token identity or syn…
On the information rate of phones, syllables and words
Which linguistic units minimise the bitrate required to encode speech content: phones, syllables, words, maybe larger units, or something in between? This question has implications for the development of speech and langu…
Specify Privacy Yourself: Assessing Inference-Time Personalized Privacy Preservation Ability of Large Vision-Language Models
Large Vision-Language Models (LVLMs) have demonstrated remarkable capabilities but raise significant privacy concerns due to their abilities to infer sensitive personal information from images with high precision. While…
Coding Agent Memory Post-Training II: Unlocking the Memory Potential of Pre-trained File Operations for Long-Horizon Tasks via On-Policy Distillation
Language-model agents increasingly tackle long-horizon tasks whose interaction histories exceed the model's active context. They must therefore decide what to write down and what to read back, and when to do either. Rece…
One Sequence Is Enough: Fingerprinting Language Models from Fixed-Length Integer Sequences
Users of model APIs and third-party platforms often rely on the declared model name to identify the model responding to their requests. However, the actual model may differ from the declared one, motivating identity veri…
Program-Aided Patchable Skill Evolution
Agent skills have emerged as an effective means of enabling LLM agents to solve complex, interactive tasks by encoding reusable knowledge without modifying model parameters. Recent approaches further improve skill adapta…
MISO-BR: An Open Corpus for Misogyny Detection in Brazilian Portuguese
Misogyny is among the most prevalent forms of online hate speech directed at Brazilian users, yet open and auditable resources dedicated to it in Brazilian Portuguese remain scarce. We introduce MISO-BR, an open, documen…
Thermodynamics of words
We introduce the word thermalization hypothesis according to which in a literary work there are effective interactions between words that drive them to thermalization where their word frequency distribution is the same a…
HandsOnAid: Building and Evaluating an Offline First-Aid Assistant for Emergency Response Scenarios
Emergencies can occur where there is no phone signal and no ambulance nearby. A small language model on a phone can be helpful in such situations, but its advice needs to be safe and accurate. We built HandsOnAid, a 4-bi…
From Bias to Positionality: A critical grammar for Large Language Models
Bias has become a default lens for understanding what machine learning systems do and how we should relate to them. Bias is at root a deviation from a standard; in decision making systems we can supply a standard from ou…
Sckwoky at SHROOM-Visions: Character-Level Hallucination-Span Detection in Image-Conditioned VLMs via Dual-Prompted LoRA Decoders
We present our system for the SHROOM-visions 2026 shared task on character-level hallucination span detection in image-conditioned vision-language model (VLM) responses, evaluated across English, French, Italian, and Chi…
Visual Similarity in Adult Word Extension and Artificial Word Learning
When two objects share a name, how much does their visual similarity explain, and what do the teaching examples add? First, we reanalyzed trial-level data from a large adult word-extension study (Lewis & Frank, 2018), co…
Representational Atoms in Neural Networks
Understanding the internal representations of neural networks and characterizing their relationship with human conceptual structure are two critical questions for elucidating model mechanisms. The former reveals how lear…
Breakdown of Autoregressive Multi-Step Reasoning Under Concentrated vs. Diffuse Attention Head Dropout
During extended wakefulness, mammalian neocortical networks experience localized periods of transient neuronal silence termed off-states. Here, we examine how analogous spatiotemporally concentrated disruptions affect se…
Stagewise and interpretable emergence of inside-outside computation in Transformer-based language models
Variability in syntactic structure and parsing ambiguity are fundamental aspects of natural language which modern LLMs have to learn to deal with proficiently. Previous "post-mortem" studies have shown that fully trained…
StaRSE: Compact Russian Sentence Embeddings with a Sign-Coded Static Encoder
Transformer sentence encoders provide strong retrieval quality, but their inference cost becomes a bottleneck when embeddings must be produced on CPUs or refreshed frequently. We introduce StaRSE, a compact Russian stati…
Dyship: Leveraging Acoustic Features For Speaker Relationship Classification From Dyadic Conversations
We introduce DyShip, a framework for classifying speaker familiarity and kinship from dyadic conversations in the Seamless Interaction corpus. We compare five speech representations: spectrograms, MFCCs, self-supervised…
An AI Assistant for Automating Anti-Corruption Expertise of Draft Legislation Based on the TRACE Framework
We present an AI assistant for automating anti-corruption expertise of draft legislation based on the TRACE architecture, designed to screen bills for corruption risks: pinpoint provisions an expert would criticise and n…
Semantic Chains: From Reachable Facts to Justified Answers
How can the facts that justify an answer be found from a question and presented as a sequence of connected statements? This paper models a context as a graph of logical facts and takes as given which expressions represen…
MMDM: Measuring Massive Multitask Decision Making
System-1 Decision Models directly make structured probabilistic decisions over query-specified options given a state in a single forward pass, without autoregressively generating reasoning trajectories or textual respons…
Point-in-Time Encoders for Earnings Calls
Text encoders turn earnings calls into vectors for search, for language models that answer from retrieved passages, and for forecasting models. A live retriever can use an encoder trained once on all available text, as f…
AI Isn't Getting Dumber, We're Getting Pickier: Rising Standards in Human Evaluation of Large Language Models
Users often complain that AI assistants such as ChatGPT are getting worse. Some versions do change, but human evaluation leaves a second possibility untested: the people judging may change. We separate the two on Chatbot…
EngineAgent: Agentizing On-Device LLM Inference Engines with Inference Stack Hierarchy Awareness and Hardware Abilities
Rapidly evolving LLM architectures and heterogeneous mobile hardware make supporting new models in on-device inference engines a recurring and costly system challenge. Each model may require coordinated changes across co…
ARE-bench: Benchmarking Agents' Research Ability in Real-World Researcher Sessions and Workspaces
Research agents achieve increasingly strong benchmark scores, yet these scores do not necessarily translate into reliable delivery in ongoing research projects. Many research-agent benchmarks emphasize standalone tasks a…
Blurring the Block Boundaries: Overlapping Block Denoising for Efficient Diffusion Language Models
Diffusion Language Models (DLMs) have emerged as effective alternatives to autoregressive models for latency-critical tasks, enabling faster token generation. Recent advances employ block diffusion with self-editing capa…
REAL Loss: Optimizing Real Expected Acceptance Length for Speculative Decoding
Speculative decoding accelerates LLM inference by having a draft model propose multiple tokens for the target model to verify in one pass. The speedup is governed by the expected acceptance length (EAL), yet standard tra…
Bits or permutations? Spending a fixed inference budget on weight precision or on option-order averaging in small local language models
Local language models can save memory by storing weights in fewer bits, and they can reduce sensitivity to the order of answer options by averaging a few reorderings. We ask whether the first makes the second worse, and…
Earlybird: Predictive Streaming Speech Recognition in 29M Parameters
Streaming speech recognizers output a word only after the speaker has finished it, typically 90-400 ms later. For a voice agent every millisecond counts: the sooner the words are outputted, the sooner the agent can start…
When does text encoding matter for language modeling?
UTF-8's role as the de facto encoding standard in language models introduces cross-lingual inequities due to its variable-length design. However, the effect that this encoding choice has on model performance remains larg…