prompt
stringlengths
41
511
target
stringlengths
1
25
keyword
stringclasses
697 values
full_sentence
stringlengths
48
1.25k
current steering methods such as vector addition and directional ablation are constrained within a two-dimensional subspace defined by the activation and feature direction making them sensitive to chosen parameters and potentially affecting unrelated features due to unintended interactions in
activation
angular steering
current steering methods such as vector addition and directional ablation are constrained within a two-dimensional subspace defined by the activation and feature direction making them sensitive to chosen parameters and potentially affecting unrelated features due to unintended interactions in activation space
large language model llm -powered chatbots are
increasingly
large language
large language model llm -powered chatbots are increasingly used for opinion exploration
here we present a systematic analysis of four
quasars
galactic disk
here we present a systematic analysis of four quasars initially selected by their ks-band variability amplitudes in the vista variables in the v i a l actea survey vvv vvvx
recent large language model llm research has undergone an architectural shift from encoder-decoder
modeling
large language models llms
recent large language model llm research has undergone an architectural shift from encoder-decoder modeling to nowadays the dominant decoder-only modeling
we outline a multi-phase research methodology focused on preparing re datasets fine-tuning ai models and designing collaborative
human-ai
human-ai interaction
we outline a multi-phase research methodology focused on preparing re datasets fine-tuning ai models and designing collaborative human-ai workflows
such algorithms are based on the minimization of a
certain
efficiently solving
such algorithms are based on the minimization of a certain cost functional
several researchers introduced different methods to decompose n 1 -bit toffoli gates in a quantum circuit into a set of standard 3-bit toffoli gates or a set of elementary quantum
gates
toffoli gates
several researchers introduced different methods to decompose n 1 -bit toffoli gates in a quantum circuit into a set of standard 3-bit toffoli gates or a set of elementary quantum gates such as single-qubit and two-qubit gates
in this work we aim to explain this conflict by exploring how
language
language models
in this work we aim to explain this conflict by exploring how language models manipulate numbers and quantify the lower bounds of accuracy of these mechanisms
our analysis allows for estimating a diverging number of
treatment
average treatment effect
our analysis allows for estimating a diverging number of treatment effects simultaneously and establishes the consistency and asymptotic normality of the regression-based estimators
will this trend towards greater concentration in large
cities
large cities
will this trend towards greater concentration in large cities continue or level off
in this paper we propose adasdbo a fully problem-parameter-free algorithm for decentralized
bilevel
quadratic programming
in this paper we propose adasdbo a fully problem-parameter-free algorithm for decentralized bilevel optimization with a single-loop structure
by applying both exact diagonalization and density-matrix renormalization group theory to the
quantum
quantum technologies
by applying both exact diagonalization and density-matrix renormalization group theory to the quantum heisenberg hamiltonian we show how a quantum-mechanical treatment of an ab initio parametrized spin model can significantly improve the predicted low-temperature spin-flop field over a classical description when compared to measurements
potential thermal profiles of the third interstellar
object
interstellar medium
potential thermal profiles of the third interstellar object 3i atlas
this design incorporates a temporal attention in vision encoder enabling the model to better capture the progression of
actions
action recognition
this design incorporates a temporal attention in vision encoder enabling the model to better capture the progression of actions and the relationships between frames before passing visual tokens to the llm
our results advance and offer new insights into the interplay between
ferrers
ferrers dimension
our results advance and offer new insights into the interplay between ferrers dimensions and extremal combinatorics
third traditional difference-in-differences methods that ignore
spatial
spatial treatment
third traditional difference-in-differences methods that ignore spatial and network structure exhibit 61 percent bias in estimated treatment effects
combined with the distribution of serial intervals or generation times the rate gives basic and instantaneous values of the
reproduction
basic reproduction
combined with the distribution of serial intervals or generation times the rate gives basic and instantaneous values of the reproduction number that govern development and ultimate outcome of the epidemic
we compare these models in terms of theoretical properties optimization strategies and empirical
performance
deep learning
we compare these models in terms of theoretical properties optimization strategies and empirical performance and discuss applications in fields such as computer vision natural language processing and bioinformatics
this paper starts by providing an overview of the
estimators
ate estimation
this paper starts by providing an overview of the estimators computed by the command
the 4 -dtc phases appear for certain initial states that are close to the spiral saddle
points
-dtc phases
the 4 -dtc phases appear for certain initial states that are close to the spiral saddle points identified in the classical picture
then we develop a collection of safety margin results on the overall control
u
control law
then we develop a collection of safety margin results on the overall control u u0 u
therefore the only scalable and viable solution to solve anomaly detection over very different time series collected from diverse domains is to propose a model selection method that will select based on time
series
time series classification
therefore the only scalable and viable solution to solve anomaly detection over very different time series collected from diverse domains is to propose a model selection method that will select based on time series characteristics the best anomaly detection methods to run
ais have made rapid progress on research-oriented benchmarks of knowledge and
reasoning
reasoning capabilities
ais have made rapid progress on research-oriented benchmarks of knowledge and reasoning but it remains unclear how these gains translate into economic value and automation
we quantify the dependence of magnetic fields on star formation
activity
star formation
we quantify the dependence of magnetic fields on star formation activity including both regular and starburst galaxies
since the gradient of the objective function is inaccessible as a result of the unknown distribution various
zeroth-order
first-order methods
since the gradient of the objective function is inaccessible as a result of the unknown distribution various zeroth-order methods have been developed to solve the problem
we employ in-game performance scores as quantitative
metrics
superior performance
we employ in-game performance scores as quantitative metrics to assess performance across different task types
carbon-aware optimal power flow with data-driven
carbon
power flow
carbon-aware optimal power flow with data-driven carbon emission tracing
instanton theory has arisen as a practical tool for calculating tunneling splittings in
molecular
instanton theory
instanton theory has arisen as a practical tool for calculating tunneling splittings in molecular systems
however existing models primarily focus on general-purpose code
review
code review
however existing models primarily focus on general-purpose code review their effectiveness in identifying and addressing security-related issues remains underexplored
in this work we investigate whether large language models llms can translate human natural
language
vision-language models
in this work we investigate whether large language models llms can translate human natural language instructions into the internal symbolic representations that emerge during hierarchical reinforcement learning
scale invariance and statistical significance in complex
weighted
correlation network
scale invariance and statistical significance in complex weighted networks
however in data-sensitive fields such as healthcare the lack of high-quality domain-specific training corpus hinders llms
adaptation
llm post-training
however in data-sensitive fields such as healthcare the lack of high-quality domain-specific training corpus hinders llms adaptation for specialized applications
our method outperforms state-of-the-art large-scale
eeg
brain-computer interface
our method outperforms state-of-the-art large-scale eeg models by an average of 4
the complexity bounds we obtain are confirmed numerically for a specific case with the o n 2 quantum algorithm outperforming the classical greedy algorithm that also runs in o
n
open quantum
the complexity bounds we obtain are confirmed numerically for a specific case with the o n 2 quantum algorithm outperforming the classical greedy algorithm that also runs in o n 2
hence great effort has been put into identifying subclasses of integer programs that are solvable in
polynomial
-time algorithm
hence great effort has been put into identifying subclasses of integer programs that are solvable in polynomial or mathsf fpt time
the growing complexity of integrated photonics necessitates compact low-power
devices
photonic devices
the growing complexity of integrated photonics necessitates compact low-power devices that transcend traditional material-centric design approaches
we resolve this question by proving a sharp
lower
time complexity
we resolve this question by proving a sharp lower bound of n log varepsilon -1 n log e - o n bits for varepsilon o 1 regardless of operation time
one such family consists of all codes with up to three
maximal
neural codes
one such family consists of all codes with up to three maximal codewords
notably under a similar performance guarantee as in our
tree
spanning trees
notably under a similar performance guarantee as in our tree embedding algorithms i
the inference cost of large language models
llms
models llms
the inference cost of large language models llms has become a critical factor in determining their commercial viability and widespread adoption
r3 adaptively samples training experience from diverse intersections with environment feedback-based priority and fine-tunes
llm
llm agents
r3 adaptively samples training experience from diverse intersections with environment feedback-based priority and fine-tunes llm agents with a designed reward-weighted likelihood loss guiding reg-tsc toward high-reward policies across heterogeneous intersections
this result generalizes existing tractability results for special
sparse
regular graphs
this result generalizes existing tractability results for special sparse families such as necklace graphs
despite multiple efforts to understand the problem in these simple graph families the computational complexity of the problem had remained unsettled and our hardness results answer open questions by bhabak and harutyunyan and harutyunyan and hovhannisyan concerning the problem s complexity in k -cycle and k -path
graphs
regular graphs
despite multiple efforts to understand the problem in these simple graph families the computational complexity of the problem had remained unsettled and our hardness results answer open questions by bhabak and harutyunyan and harutyunyan and hovhannisyan concerning the problem s complexity in k -cycle and k -path graphs respectively
while large language models llms offer opportunities in document
understanding
vision-language models
while large language models llms offer opportunities in document understanding current systems struggle with complex multi-page visual documents particularly in fine-grained reasoning over elements and pages
6d channel knowledge map construction via bidirectional
wireless
channel state information
6d channel knowledge map construction via bidirectional wireless gaussian splatting
multiplexes have emerged as a key instrument for modeling large-scale complex
systems
complex systems
multiplexes have emerged as a key instrument for modeling large-scale complex systems due to the widespread coexistence of diverse interactions in social industrial and biological domains
this paper proposes a novel haoi sequence generation framework synhlma to synthesize hand
language
vision-language-action vla
this paper proposes a novel haoi sequence generation framework synhlma to synthesize hand language manipulation for articulated objects
objective to develop validate and deploy a
natural
natural language processing
objective to develop validate and deploy a natural language processing nlp pipeline to identify itfs in radiology reports and assess their prevalence features and clinical outcomes
we propose a verification-informed nn framework that incorporates worst-case constraint
violations
soft constraints
we propose a verification-informed nn framework that incorporates worst-case constraint violations directly into training producing models that are both accurate and provably safer
the use of ad-hoc engineered viruses in the fight against
tumours
viral replication
the use of ad-hoc engineered viruses in the fight against tumours is one of the greatest ideas in cancer therapeutics within the last three decades
in time series modeling achieving a balance between model performance and computational efficiency
remains
time series
in time series modeling achieving a balance between model performance and computational efficiency remains a significant challenge
semantic representations emerge in biologically
inspired
artificial neural
semantic representations emerge in biologically inspired ensembles of cross-supervising neural networks
these findings suggest that reinforcement
learning
reinforcement learning
these findings suggest that reinforcement learning approaches can be effectively adapted for use in random nonstationary and reward-sparse environments
these accretion-based m_ bullet values which in this sample span nearly three orders of magnitude are consistent with
galactic
stellar mass
these accretion-based m_ bullet values which in this sample span nearly three orders of magnitude are consistent with galactic scaling relations but are significantly more precise 68 credible interval pm 0
our framework is particularly applicable to non-homogeneous human populations offering a new perspective on cooperation science in the context of cultural evolution where neutral and biased processes within structured
interactions
social interactions
our framework is particularly applicable to non-homogeneous human populations offering a new perspective on cooperation science in the context of cultural evolution where neutral and biased processes within structured interactions are abundant
by choosing their particular case we systematically study the impact of these
interactions
quantum walk
by choosing their particular case we systematically study the impact of these interactions on the dynamics of two initially localized and noncorrelated walkers
this paper introduces a quantitative economics of inference framework treating the llm
inference
moe inference
this paper introduces a quantitative economics of inference framework treating the llm inference process as a compute-driven intelligent production activity
we address these issues with graph-enhanced policy optimization gepo which dynamically constructs a state-transition graph from agent experience and employs graph-theoretic centrality to provide three synergistic
learning
reinforcement learning
we address these issues with graph-enhanced policy optimization gepo which dynamically constructs a state-transition graph from agent experience and employs graph-theoretic centrality to provide three synergistic learning signals 1 structured intrinsic rewards that guide exploration toward high-impact states 2 a graph-enhanced advantage function for topology-aware credit assignment and 3 a dynamic discount factor adapted to each state s strategic value
finally edges are ranked depending on the observed gromov-wasserstein distance the higher the value of the
distance
gromov-wasserstein distance
finally edges are ranked depending on the observed gromov-wasserstein distance the higher the value of the distance the more critical the edge is in terms
these patterns often follow power- law distributions in inter-event intervals or
event
statistical physics
these patterns often follow power- law distributions in inter-event intervals or event rates
chirped control pulses designed to satisfy
adiabatic
coherent control
chirped control pulses designed to satisfy adiabatic conditions across the ensemble enable broadband rephasing
we compare these models in terms of theoretical properties optimization strategies and empirical performance and discuss applications in fields such as computer vision natural
language
language models
we compare these models in terms of theoretical properties optimization strategies and empirical performance and discuss applications in fields such as computer vision natural language processing and bioinformatics
these formulations also allow us to arrive at better analytical characterisation of how waves of
viral
viral replication
these formulations also allow us to arrive at better analytical characterisation of how waves of viral infections arise and propagate in tumours providing interesting insights into therapy dynamics
cypress leverages a pre-trained large-scale geospatial
foundation
foundation models
cypress leverages a pre-trained large-scale geospatial foundation model prithvi-eo-2
based on this reformulation we propose a single-loop first-order algorithm for linearly constrained
bilevel
optimization problem
based on this reformulation we propose a single-loop first-order algorithm for linearly constrained bilevel optimization sflcb
in this work we study the limits of compressed
data
data structure
in this work we study the limits of compressed data structures i
fractalbrain presents an experience combining a surreal
virtual
physical virtual
fractalbrain presents an experience combining a surreal virtual reality vr program with an electroencephalogram eeg interface
our analysis allows for estimating a diverging number of
treatment
dynamic treatment
our analysis allows for estimating a diverging number of treatment effects simultaneously and establishes the consistency and asymptotic normality of the regression-based estimators
effectiveness and user behavior were assessed through a controlled experiment in which participants interacted with either persona while a control group engaged with a standard
llm
llm responses
effectiveness and user behavior were assessed through a controlled experiment in which participants interacted with either persona while a control group engaged with a standard llm providing direct answers
unlike headset-based vr textit storycaster preserves spatial awareness using live camera feeds to augment the walls with cylindrical projections allowing users to create worlds that blend with their
physical
virtual reality
unlike headset-based vr textit storycaster preserves spatial awareness using live camera feeds to augment the walls with cylindrical projections allowing users to create worlds that blend with their physical surroundings
the local gaussian correlation coefficient a new measurement of statistical dependence between variables has unique advantages including capturing
local
local gaussian
the local gaussian correlation coefficient a new measurement of statistical dependence between variables has unique advantages including capturing local nonlinear dependence and handling heavy-tailed distributions
we also discuss potential limitations including completeness impacts of background subtraction and spatial
resolution
galaxy cgm
we also discuss potential limitations including completeness impacts of background subtraction and spatial resolution and the generality of our results when applied to other galaxies
designing efficient quantum circuits is a central bottleneck to exploring the potential of quantum
computing
quantum error correction
designing efficient quantum circuits is a central bottleneck to exploring the potential of quantum computing particularly for noisy intermediate-scale quantum nisq devices where circuit efficiency and resilience to errors are paramount
interestingly for the trapped system a cat-like superposition state corresponds to maximum entanglement
entropy
quantum correlations
interestingly for the trapped system a cat-like superposition state corresponds to maximum entanglement entropy below the transition highlighting the relevance of the present model for studying the effect of decoherence on intra-particle entanglement in the context of quantum information processing
by finite-size scaling of the monte carlo data we identified the critical exponents of the magnetization
magnetic
magnetic properties
by finite-size scaling of the monte carlo data we identified the critical exponents of the magnetization magnetic susceptibility and spin correlation length beta 0
the central density of hii regions correlates well with the surface brightness within the
effective
hii regions
the central density of hii regions correlates well with the surface brightness within the effective radius
see4d pose-free 4d generation via auto-regressive
video
image generation
see4d pose-free 4d generation via auto-regressive video inpainting
here we show that lyman- alpha ly alpha radiation pressure limits the gas mass converted into
stars
massive stars
here we show that lyman- alpha ly alpha radiation pressure limits the gas mass converted into stars particularly in primordial environments
we propose acc-sgd-ie a trajectory-aware estimator that propagates the leave-one-out perturbation across
training
continual learning
we propose acc-sgd-ie a trajectory-aware estimator that propagates the leave-one-out perturbation across training and updates an accumulative influence state at each step
besides ac conductivity electric modulus and dielectric
properties
electronic structure
besides ac conductivity electric modulus and dielectric properties have been investigated to illustrate the microscopic conduction mechanism
dataset bias where data points are skewed to certain concepts is ubiquitous in
machine
machine learning
dataset bias where data points are skewed to certain concepts is ubiquitous in machine learning datasets
l_ bol disk l_ edd propto t_p 4 propto m_ bullet -1 tde-specific physics correlations l_ plat propto m_ bullet 2 3 and r_ out r_g propto m_ bullet -2 3 and black hole-host galaxy correlations m_ bullet - m_ star and m_ bullet - sigma_ star naturally emerge from the data and for the first time are self-consistently extended into the intermediate-mass imbh m_
bullet
stellar mass function
l_ bol disk l_ edd propto t_p 4 propto m_ bullet -1 tde-specific physics correlations l_ plat propto m_ bullet 2 3 and r_ out r_g propto m_ bullet -2 3 and black hole-host galaxy correlations m_ bullet - m_ star and m_ bullet - sigma_ star naturally emerge from the data and for the first time are self-consistently extended into the intermediate-mass imbh m_ bullet 10 5 regime
first we show that collective bias is a deeper phenomenon than previously assessed interaction can amplify
individual
collective action
first we show that collective bias is a deeper phenomenon than previously assessed interaction can amplify individual biases introduce new ones or override model-level preferences
however this paper demonstrates that a minimal correction applied solely to the upper bound iterate is sufficient to boost the
convergence
convergence rate
however this paper demonstrates that a minimal correction applied solely to the upper bound iterate is sufficient to boost the convergence of the method achieving superquadratic and cubic rates for the upper and lower bound sequences respectively
in this paper we train a neural network nn to solve the optimal
power
optimal power flow
in this paper we train a neural network nn to solve the optimal power shutoff line switching problem
meanwhile it provides a unified theoretical explanation for classic topological characteristics such as small-world networks and
scale-free
complex networks
meanwhile it provides a unified theoretical explanation for classic topological characteristics such as small-world networks and scale-free networks
this paper presents a fully data-driven control framework for autonomous underwater vehicles auvs based on data-enabled
predictive
predictive control
this paper presents a fully data-driven control framework for autonomous underwater vehicles auvs based on data-enabled predictive control deepc
traditional machine learning relies on explicit models and domain assumptions limiting
flexibility
machine learning
traditional machine learning relies on explicit models and domain assumptions limiting flexibility and interpretability
our analysis reveals two critical limitations of existing
video-llms
video generation
our analysis reveals two critical limitations of existing video-llms 1 models often fail to maintain consistency with results far worse than their single-view performances
this framework can be applied to policy evaluation using the panel data approach pda where we further establish inference for the
average
policy evaluation
this framework can be applied to policy evaluation using the panel data approach pda where we further establish inference for the average treatment effect
the grammar introduces new composable elements that support visualization across linear cyclical quasi-cyclical and other granularities standardization of irregular durations and alignment of time points across different granularities and
time
temporal semantics
the grammar introduces new composable elements that support visualization across linear cyclical quasi-cyclical and other granularities standardization of irregular durations and alignment of time points across different granularities and time zones
this paper develops a nonparametric framework for identifying and estimating
spatial
treatment effect boundaries
this paper develops a nonparametric framework for identifying and estimating spatial boundaries of treatment effects in settings with geographic spillovers
2 less than humans highlighting the potential for enabling efficient
collaboration
human-machine teaming
2 less than humans highlighting the potential for enabling efficient collaboration by delegating easily programmable tasks to agents
energy approach from varepsilon -graph to continuum
diffusion
diffusion models
energy approach from varepsilon -graph to continuum diffusion model with connectivity functional
understanding how individual learning behavior and structural
dynamics
opinion dynamics
understanding how individual learning behavior and structural dynamics interact is essential to modeling emergent phenomena in socioeconomic networks
2 we design an algorithm that gathers all of them in
poly
time complexity
2 we design an algorithm that gathers all of them in poly n log lambda time where n resp
recent advances advocate easily obtainable
channel
channel state information
recent advances advocate easily obtainable channel state information csi by commercial wifi devices for lightweight rf fingerprinting while falling short in addressing the challenges of coarse granularity of csi measurements in an open-world setting
quantum error correction qec protects qubits against bit- and phase-flip errors in the 0 or 1 subspace but physical
qubits
qubit readout
quantum error correction qec protects qubits against bit- and phase-flip errors in the 0 or 1 subspace but physical qubits can also leak into higher energy levels e
specifically we design an in-context conditioning strategy that
prompts
language models
specifically we design an in-context conditioning strategy that prompts the model with a reference example
in many problems involving causal effects or structural models the parameters of interest depend on
regression
regression function
in many problems involving causal effects or structural models the parameters of interest depend on regression functions