Category
2,491 papers
Evolutionary Soups, a mixture-of-experts framework for fine-grained generation control, with gating networks trained via an evolutionary algorithm, achieves the best hypervolume, linear utility, and Tchebyshev utility among controllable methods on all tasks.
This work presents a systematic mechanistic analysis of multilingual safety using sparse autoencoder features, sparse interpretable directions in the residual stream associated with harmful and harmless model behavior across three instruction-tuned LLMs, eight languages, and all model layers to qualify the language-universality of safety alignment as architecture-dependent and offer a mechanistic account of multilingual safety interventions.
This is the first work to formulate compression-aware abstention as a learning problem, in which a model learns to answer when supporting evidence survives compression and abstain when it does not, and controlled-deletion experiments show that the learned behavior is driven by evidence content rather than input length alone.
Experiments show Influence-Directed Adaptive On-Policy Distillation (IDA-OPD), rather than relying on costly full-vocabulary Forward-KL objectives, preserves entropy-expanding updates while replacing entropy-contracting ones with divergence-adaptive advantage shrinkage, using only the teacher's sampled-token log-probability.
Reach audiences
Advertise in front of researchers, engineers, and readers.
This work examines whether FVs extracted from a source language can steer task behavior in another language under both standard clean and perturbed zero-shot settings without providing demonstrations during inference, and observes that each LLM exhibits a relatively stable optimal range of attention heads for constructing effective FVs, and the pattern remains consistent across languages.
Gradient-based jailbreak suffix optimization methods typically update the suffix by retaining the candidate with the lowest current loss. We show that this seemingly natural design is fundamentally myopic: candidates that look better under the current-step proxy often fail to produce better jailbreak outcomes later in the search, revealing a form of selection-stage reward hacking. This suggests that candidate selection, rather than candidate generation alone, is a hidden bottleneck in suffix optimization. To address this issue, we propose \OURS{}, a trajectory-aware candidate selection framework for jailbreak suffix optimization. Instead of selecting candidates solely by their immediate loss, \OURS{} augments per-step evaluation with a trajectory-aware proxy and stabilizes selection with reference-policy regularization and a discriminator-estimated chi-squared correction, encouraging choices that remain effective beyond the current step. Experiments on HarmBench show that \OURS{} consistently outperforms strong baselines under the same search budget, substantially improving attack success rates while exhibiting more stable optimization behavior throughout the search. Our findings highlight that mitigating selection-stage reward hacking caused by myopic candidate selection is critical for improving jailbreak suffix optimization.
As context lengths scale, attention increasingly becomes a primary computational bottleneck in large language models. Standard Transformers remain powerful but computationally inefficient, as they allocate the same attention budget to every token regardless of its contextual demand. Existing local-global hybrids provide a more efficient alternative by mixing restricted- and full-context attention, but they typically allocate span statically across layers or heads. To address these limitations, we propose LoGo, a token-level dynamic local-global attention mechanism that uses attention span as a direct proxy for attention budget allocation. Each LoGo layer contains coupled local and global branches: all tokens receive efficient local attention over a restricted context window, while a learned gate activates global attention with full-context access only for tokens requiring long-range information. A threshold-based budget controller maintains a target global ratio without auxiliary losses, and a progressive masking schedule stabilizes training before sparse routing takes effect. We further implement query-sparse Triton kernels that convert reduced global-attention computation into practical speedups. Extensive experiments validate LoGo's effectiveness, showing that it preserves the scaling behavior of full-attention Transformers across model sizes. In controlled comparisons, LoGo improves over the full-attention Transformer and matched-budget static local-global hybrids, with clear gains on long-range retrieval. Analysis further shows that LoGo learns interpretable span allocation patterns. These results suggest that learned token-level span allocation is an effective and scalable way to improve the long-context performance-compute trade-off.
A benchmark score is a joint property of the model, the evaluation harness, the elicitation budget, the sampled population, and contamination status. Leaderboards publish the model and the score, so capability and leakage stay observationally equivalent. Existing taxonomies classify contamination for automated detection, not the question a reporter faces at publication: given the mitigations already applied, which validity threats remain open? We introduce a taxonomy organized by the mitigation each type defeats -- direct, derivative, temporal, distributional, and acquired -- spanning training-time and evaluation-time leakage. Holding out a private test set closes the first alone. The fifth is acquired during the evaluation itself; because it is a property of one run, it must be recorded with the reported score rather than with the benchmark release. We operationalize it as a four-field disclosure protocol in which"unknown"is a valid entry, released under CC BY 4.0 with a JSON Schema, a validator, and worked examples. Two coders external to the design team applied a pre-registered instrument to 41 documents. Per-variable linear-weighted $\kappa$ runs from 0.00 to 0.35 (median 0.21) over 29 main-pass documents against a single-coder test-retest ceiling of 0.84, collapsing under the class skew the registration anticipated; pooling raises it to 0.46 through chance correction rather than better agreement. Two variables fall below the prevalence-robust threshold registered in advance: strata reporting and the acquired type introduced here. Disagreement concentrates on when a variable applies rather than on what a document states. Elicitation budgets are reported in 13% of documents, and no document addresses all five types. The contribution is the taxonomy, the score-side artifact that follows from it, and a pre-registered measurement of instrument reliability and current disclosure.
LLMs are increasingly used as human surrogates, often on the premise that richer persona data could make them substitutes or exploratory tools for specific individuals. We test this premise across four datasets covering more than 400,000 participants and more than 6,000 survey items and experimental outcomes. LLMs perform well at the aggregate level: their average responses closely align with average human responses to the same items. But this success largely reflects predicting each item's average human response. Once each item's human mean is removed, LLM predictions explain only 3.05% of the remaining respondent-specific variation, far below the 53.6% human test-retest benchmark. Richer personas, model variants, and fine-tuning do not close this gap. In variance analyses, once item means are removed, the reliable remaining signal is person-by-item. It captures how a respondent departs from the mean on a particular item and is about 8.9x larger than the stable person effect. Persona data encode the respondent, but not this item-specific deviation. LLM responses also compress human response distributions, using less spread, fewer response categories, and distorted distributional shapes. We call this pattern item-mean surrogacy. Current LLM surrogates can approximate item averages, but not the distributions or respondent-specific deviations needed to replace individual humans. We propose four empirical tests for LLM-based human-surrogate claims.
This work proposes that LLM web agents can learn simple environment observations at test time, and introduces trial steps for agents to decompose a complex environment observation into sub-modules, and implements a label-free learning method, Test-Time Environment Decomposition (TTED), to adapt agent behaviors with experience during inference.
This paper provides a practical, auditable, and application-agnostic framework for validating LLM-augmented recipe data, thereby strengthening the foundations of machine-readable food knowledge infrastructures in the era of LLM-generated content.
The EM generalization metric is extended from a scalar distance to a dataset-specific generalization direction, which robustly predicts EM models'evilness under semantics-preserving prompt perturbations including appending random tokens and paraphrasing, where other methods do not reliably generalize.
A new method, called CW-Net, translates the reasoning process of an autonomous vehicle’s AI system into understandable concepts that explain its behavior.
With millions of users across the world, Julia has been used to conduct cutting-edge research and to design new drugs, jet engines, heat pumps, and more.
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
A new machine-learning framework aims to improve the success rate of computational protein design while moving away from results that reproduce sequences found in nature.
Algorithms & Theory
New MIT research could lead to better materials for a fossil-fuel-free process for making the chemical that's essential to fertilizer and other products.
We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.
