MLINFO COLLECTION
LoRA、RLHF、SFT、軽量ファインチューニングなど、LLMを用途に合わせる技術を扱います。
AI assistants built on large language models (LLMs) are trained on human feedback and increasingly personalized to users' stated preferences. Yet how closely these preferences reflect what people…
This paper presents AgentDP, which aims to deploy pre-trained AI models on heterogeneous embedded devices by fully automating model optimization and adaptation (O&A) with Large Language Model (LLM)…
Masked diffusion models (MDMs) generate sequences by progressively unmasking tokens, but errors introduced during decoding are typically irreversible. Recent work has shown that iterative correction…
Guardrail models are increasingly used as safety layers for LLMs and agentic systems, but most are trained on fixed prompt or response interfaces and become unreliable when the same safety case…
Shadows are observable consequences of latent physical causes, arising from the interaction between illumination, object geometry, and scene structure. A model that understands shadows should be able…
Large language models have been reported to encode the harmfulness of a request separately from their decision to refuse it, and latent safety monitors such as Latent Guard build on this report by…
This paper studies the local geometry of semantic meaning in large language models. Current representation analyses ask whether meanings are close in embedding space, or whether linear probes…
Text embedding and text generation are two of the most useful capabilities of modern large language models, yet they are typically provided by separate specialized models (an embedder and a…
An LLM agent can retrieve a memory correctly yet use it incorrectly after its validity state changes. We study governance-state corruption: faults that alter which retained records enter…
When a multilingual LLM agrees with a user's false premise, is it being sycophantic, or does it simply not know better in that language? Current research does not distinguish these cases: they record…
Low-rank adaptation (LoRA) is usually tied to the pretrained backbone on which it is learned. We ask whether LoRAs trained independently for the same task on different pretrained backbones can be…
Large language models (LLMs) deployed in mental-health contexts can be evaluated for safety failures, but the resulting judgments are often difficult to inspect or explain. Common approaches to…
Existing datasets on cultural knowledge have been typically created by human annotators familiar with each culture. However, the process requires considerable human effort, presenting a significant…
Large language models increasingly act as controllers in structured decision pipelines, yet their decisions are still executed by autoregressive text decoding. Hierarchical RAG is one such setting:…
Autoregressive encoder-decoder speech recognition models trained with teacher forcing can suffer from exposure bias, limiting their ability to recover from their own prediction errors at inference…
Class labels offer compact supervision for describing cultural-heritage images when expert text annotations are scarce. We investigate a batik-captioning cascade that integrates a LoRA-adapted DINOv2…
Social world models predict how interactions develop after an action, helping agents negotiate, cooperate and mediate. In these situations, one action can lead to several plausible futures, hence…
The claim that speculative-decoding acceptance rate decays as a chain-of-thought grows longer has become a de facto industry narrative, motivating draft-model co-design and position-aware training.…
Coding-agent benchmarks often describe a task in natural language but grade it with tests taken from the upstream project. Passing those tests can require information the description never states: a…
Medical knowledge evolves as clinical guidelines revise recommendations, yet whether large language models (LLMs) faithfully incorporate such changes remains unclear. We introduce ChronoEvo,…
Time series anomaly detection (TSAD) is a foundational problem in data mining, aiming to identify observations that deviate significantly from nominal temporal patterns. Driven by escalating data…
LLMs have shown strong in-context learning (ICL) capability, but extending it to Multimodal LLMs (MLLMs) remains challenging. Prior multimodal ICL methods often rely on ICL-specific datasets for…
AI control protocols aim to get useful work from models that may be misaligned by monitoring their behaviour. One approach uses a second copy of the same model as the monitor.
Financial news can influence investment decisions long after publication. Timely news may reach an investor before prices fully adjust, whereas older news may already be reflected in prices.