Pre-training
Collection
24 items • Updated • 1
data_source stringclasses 1
value | prompt listlengths 1 1 | ability stringclasses 1
value | reward_model dict | extra_info dict |
|---|---|---|---|---|
lexical_match_custom | [
{
"content": "Continue the generation as closely to verbatim as possible. # 1 Introduction Context awareness is essential in Human-Computer Interaction (HCI), enabling systems to detect, interpret, and respond to contextual information [3, 58, 135], thereby facilitating adaptive and seamless interactions [173].... | lexical_match | {
"ground_truth": "VMIs address unimodal limitations to deliver accurate, contextually relevant responses [74, 141]. Despite ongoing advancements in visual information processing, the use of VMIs as interactive tools remains nascent. The rapid development of technology, especially in multimodal Artificial Intelligenc... | {
"assistant_prefix": "VMIs address unimodal limitations to deliver accurate, contextually relevant responses [74, 141]. Despite ongoing advancements in visual information processing, the use of VMIs as interactive tools remains nascent. The rapid development of technology, especially in multimodal Artificial Intelli... |
lexical_match_custom | [
{
"content": "Continue the generation as closely to verbatim as possible. \\section{Introduction} Standard Model (SM) has been successful in explaining a wide variety of phenomena and most of the experimental results are consistent with the SM \\cite{Langacker:2009my}. However, there are some issues that make i... | lexical_match | {
"ground_truth": "breaking mechanism is completely unknown. For this reason, low energy SUSY spectrum can be completely arbitrary and it needs more than 100 parameters to specify the masses and couplings of the SUSY particles. Practically it is very difficult to explore such a multi-dimensional parameter space and s... | {
"assistant_prefix": "breaking mechanism is completely unknown. For this reason, low energy SUSY spectrum can be completely arbitrary and it needs more than 100 parameters to specify the masses and couplings of the SUSY particles. Practically it is very difficult to explore such a multi-dimensional parameter space a... |
lexical_match_custom | [
{
"content": "Continue the generation as closely to verbatim as possible. # 1 Introduction Multimodal document retrieval (Hassan et al., 2013; Lee et al., 2024) aims to retrieve information from visually rich documents based on user queries. Unlike traditional document retrieval (Zhang et al., 2022; Chen et al.... | lexical_match | {
"ground_truth": "text occupies only $52.7\\%$ of content area, while images and tables account for $29.2\\%$ and $12.8\\%$ respectively. This highlights the need for retrieval systems that effectively handle multimodal and cross modal (Zhang et al., 2025; Dong et al., 2025b) information. However, as shown on Table ... | {
"assistant_prefix": "text occupies only $52.7\\%$ of content area, while images and tables account for $29.2\\%$ and $12.8\\%$ respectively. This highlights the need for retrieval systems that effectively handle multimodal and cross modal (Zhang et al., 2025; Dong et al., 2025b) information. However, as shown on Ta... |
lexical_match_custom | [
{
"content": "Continue the generation as closely to verbatim as possible. \\section{Introduction} The aim of this work is to define Lagrange interpolating polynomials for a set of points and values belonging to a real Clifford algebra. We make some preliminary considerations to select the Clifford algebras on w... | lexical_match | {
"ground_truth": "denotes the $k$--vector component of $x\\in \\mathbb{R}_{p,q}$ (cf.\\ e.g.\\ \\cite[\\S4.1]{CSSSbook} or \\cite[\\S 3.2]{GHS}). For every element $x$ of $\\mathbb{R}_{p,q}$, the \\emph{trace} of $x$ is $t(x):=x+x^c$ and the (squared) \\emph{norm} of $x$ is $n(x):=xx^c$. Let $m:=p+q$. An element $x$... | {
"assistant_prefix": "denotes the $k$--vector component of $x\\in \\mathbb{R}_{p,q}$ (cf.\\ e.g.\\ \\cite[\\S4.1]{CSSSbook} or \\cite[\\S 3.2]{GHS}). For every element $x$ of $\\mathbb{R}_{p,q}$, the \\emph{trace} of $x$ is $t(x):=x+x^c$ and the (squared) \\emph{norm} of $x$ is $n(x):=xx^c$. Let $m:=p+q$. An element... |
lexical_match_custom | [
{
"content": "Continue the generation as closely to verbatim as possible. # 1 Introduction Automated evaluation of large language models (LLMs) is an increasingly pertinent task as LLMs demonstrate their value across a growing array of real-world use cases. Reliable evaluation is critical to ensure that LLMs ar... | lexical_match | {
"ground_truth": "use and easy to get started with. However, this approach poses a number of challenges. Prompted evaluations are often poorly correlated with human judgments, and addressing this requires extra time  Figure 1: Atla Selen... | {
"assistant_prefix": "use and easy to get started with. However, this approach poses a number of challenges. Prompted evaluations are often poorly correlated with human judgments, and addressing this requires extra time  Figure 1: Atla S... |
lexical_match_custom | [
{
"content": "Continue the generation as closely to verbatim as possible. # 1. Introduction Medical diagnosis of histopathology through the examination of whole slide images (WSIs) is a cornerstone of modern pathology. WSIs are high-resolution, digitally scanned histopathology cases, providing an extensive view... | lexical_match | {
"ground_truth": "deep learning report achieving expert-level performance, promising such a scalable approach [47]. However, current methods typically divide WSIs into smaller patches for independent analysis, making diagnoses without the holistic context [1, 13, 23, 26, 44, 46, 53, 54, 56]. Transformer-based models... | {
"assistant_prefix": "deep learning report achieving expert-level performance, promising such a scalable approach [47]. However, current methods typically divide WSIs into smaller patches for independent analysis, making diagnoses without the holistic context [1, 13, 23, 26, 44, 46, 53, 54, 56]. Transformer-based mo... |
lexical_match_custom | [{"content":"Continue the generation as closely to verbatim as possible. # 1. Introduction Real-worl(...TRUNCATED) | lexical_match | {"ground_truth":"like [14, 63, 67, 75] incorporate temporal blocks or optical flow maps to improve t(...TRUNCATED) | {"assistant_prefix":"like [14, 63, 67, 75] incorporate temporal blocks or optical flow maps to impro(...TRUNCATED) |
lexical_match_custom | [{"content":"Continue the generation as closely to verbatim as possible. \\section{Introduction} \\l(...TRUNCATED) | lexical_match | {"ground_truth":"be managed, or even cured, but our understanding remains poor despite impressive pr(...TRUNCATED) | {"assistant_prefix":"be managed, or even cured, but our understanding remains poor despite impressiv(...TRUNCATED) |
lexical_match_custom | [{"content":"Continue the generation as closely to verbatim as possible. \\section{Introduction} \\l(...TRUNCATED) | lexical_match | {"ground_truth":"be useful for the description of the {\\it restricted partition function} $W(s,{\\b(...TRUNCATED) | {"assistant_prefix":"be useful for the description of the {\\it restricted partition function} $W(s,(...TRUNCATED) |
lexical_match_custom | [{"content":"Continue the generation as closely to verbatim as possible. \\section{Introduction} \\l(...TRUNCATED) | lexical_match | {"ground_truth":"Over the last 50 years, there have been many proposed macroscopic mechanisms to for(...TRUNCATED) | {"assistant_prefix":"Over the last 50 years, there have been many proposed macroscopic mechanisms to(...TRUNCATED) |
End of preview. Expand in Data Studio
No dataset card yet
- Downloads last month
- 8