arXiv subjects

Jongwon Lim

Publications and source records attributed to Jongwon Lim.

8 recordsLinked to original sources

arXiv2026

ArcANE: Do Role-Playing Language Agents Stay in Character at the Right Time?

Role-playing language agents (RPLAs) simulate specific characters and personas across applications such as entertainment, companionship, interactive storytelling, and education. Faithful role-play requires more than producing plausible, in-character responses: as a character's values and behavior change over a narrative, an RPLA should reflect the character's state at the relevant stage. However, existing benchmarks largely treat characters as fixed personas or test only what they know at a given point in the narrative. We introduce ArcANE (Arc-Aware Narrative Evaluation), a benchmark for evaluating whether an RPLA follows a character's development across a narrative. ArcANE first builds an Arc that maps how a character's values, motivations, or relationships change over the story. The benchmark then scores how well an RPLA's responses fit the corresponding stages of the Arc, covering three distinct scenario types: scenes from the novel, new situations within its world, and situations outside that world. We evaluate six models under six ways of providing narrative context. In every model, using the Arc up to the queried chapter yields the best performance, outperforming the strongest non-Arc context by 2.2-8.4 points. These results suggest that faithful role-play requires evolving character states and tracking their trajectory, rather than merely retrieving relevant episodic evidence.

cs.CL↗

arXiv2026

Your Language Model is Its Own Critic: Reinforcement Learning with Value Estimation from Actor's Internal States

Reinforcement learning with verifiable rewards (RLVR) for Large Reasoning Models hinges on baseline estimation for variance reduction, but existing approaches pay a heavy price: PPO requires a policy-model scale critic, while GRPO needs multiple rollouts per prompt to keep its empirical group mean stable. We introduce Policy Optimization with Internal State Value Estimation), which obtains a baseline at negligible cost by using the policy model's internal signals already computed during the policy forward pass. A lightweight probe predicts the expected verifiable reward from the hidden states of the prompt and generated trajectory, as well as token-entropy statistics, and is trained online alongside the policy. To preserve gradient unbiasedness despite using trajectory-conditioned features, we introduce a cross-rollout construction that predicts each rollout's value from an independent rollout's internal states. Because POISE estimates prompt value using only a single rollout, it enables higher prompt diversity for a fixed compute budget during training. This reduces gradient variance for more stable learning and also eliminates the compute overhead of sampling costs for detecting zero-advantage prompts. On Qwen3-4B and DeepSeek-R1-Distill-Qwen-1.5B across math reasoning benchmarks, POISE matches DAPO while requiring less compute. Moreover, its value estimator shows similar performance to a separate LLM-scale value model and generalizes to various verifiable tasks. By leveraging the model's own internal representations, POISE enables more stable and efficient policy optimization.

cs.LG↗

arXiv2026

Learning to Retrieve User History and Generate User Profiles for Personalized Persuasiveness Prediction

Estimating the persuasiveness of messages is critical in various applications, from recommender systems to safety assessment of LLMs. While it is imperative to consider the target persuadee's characteristics, such as their values, experiences, and reasoning styles, there is currently no established systematic framework to optimize leveraging a persuadee's past activities (e.g., conversations) to the benefit of a persuasiveness prediction model. To address this problem, we propose a context-aware user profiling framework with two trainable components: a query generator that generates optimal queries to retrieve persuasion-relevant records from a user's history, and a profiler that summarizes these records into a profile to effectively inform the persuasiveness prediction model. Our evaluation on the ChangeMyView Reddit dataset shows consistent improvements over existing methods across multiple predictor models, raising F1 from 33% to 47% on Llama-3.3-70B-Instruct. Further analysis shows that effective user profiles are context-dependent and predictor-specific, rather than relying on static attributes or surface-level similarity. Together, these results highlight the importance of task-oriented, context-dependent user profiling for personalized persuasiveness prediction.

cs.CL↗

arXiv2025

Dual Mechanisms of Value Expression: Intrinsic vs. Prompted Values in Large Language Models

Large language models can express values in two main ways: (1) intrinsic expression, reflecting the model's inherent values learned during training, and (2) prompted expression, elicited by explicit prompts. Given their widespread use in value alignment, it is paramount to clearly understand their underlying mechanisms, particularly whether they mostly overlap (as one might expect) or rely on distinct mechanisms. We analyze this largely understudied problem at the mechanistic level using two approaches: (1) value vectors, feature directions representing value mechanisms extracted from the residual stream, and (2) value neurons, MLP neurons that contribute to value vectors. We demonstrate that intrinsic and prompted value mechanisms partly share common components crucial for inducing value expression, generalizing across languages and reconstructing theoretical inter-value correlations in the model's internal representations. Yet, each mechanism also possesses unique components that fulfill distinct roles. In particular, the intrinsic mechanism activates in more diverse value-related scenarios and promotes response diversity, whereas the prompted mechanism strengthens instruction compliance, taking effect even in distant tasks like jailbreaking.

cs.CL↗

arXiv2024

DAHL: Domain-specific Automated Hallucination Evaluation of Long-Form Text through a Benchmark Dataset in Biomedicine

We introduce DAHL, a benchmark dataset and automated evaluation system designed to assess hallucination in long-form text generation, specifically within the biomedical domain. Our benchmark dataset, meticulously curated from biomedical research papers, consists of 8,573 questions across 29 categories. DAHL evaluates fact-conflicting hallucinations in Large Language Models (LLMs) by deconstructing responses into atomic units, each representing a single piece of information. The accuracy of these responses is averaged to produce the DAHL Score, offering a more in-depth evaluation of hallucinations compared to previous methods that rely on multiple-choice tasks. We conduct experiments with 8 different models, finding that larger models tend to hallucinate less; however, beyond a model size of 7 to 8 billion parameters, further scaling does not significantly improve factual accuracy. The DAHL Score holds potential as an efficient alternative to human-annotated preference labels, being able to be expanded to other specialized domains. We release the dataset and code in public.

cs.CL↗

arXiv2021

Learning to increase matching efficiency in identifying additional b-jets in the $\text{t}\bar{\text{t}}\text{b}\bar{\text{b}}$ process

The $\text{t}\bar{\text{t}}\text{H}(\text{b}\bar{\text{b}})$ process is an essential channel to reveal the Higgs properties but has an irreducible background from the $\text{t}\bar{\text{t}}\text{b}\bar{\text{b}}$ process, which produces a top quark pair in association with a b quark pair. Therefore, understanding the $\text{t}\bar{\text{t}}\text{b}\bar{\text{b}}$ process is crucial for improving the sensitivity of a search for the $\text{t}\bar{\text{t}}\text{H}(\text{b}\bar{\text{b}})$ process. To this end, when measuring the differential cross-section of the $\text{t}\bar{\text{t}}\text{b}\bar{\text{b}}$ process, we need to distinguish the b-jets originated from top quark decays, and additional b-jets originated from gluon splitting. Since there are no simple identification rules, we adopt deep learning methods to learn from data to identify the additional b-jets from the $\text{t}\bar{\text{t}}\text{b}\bar{\text{b}}$ events. Specifically, by exploiting the special structure of the $\text{t}\bar{\text{t}}\text{b}\bar{\text{b}}$ event data, we propose several loss functions that can be minimized to directly increase the matching efficiency, the accuracy of identifying additional b-jets. We discuss the difference between our method and another deep learning-based approach based on binary classification arXiv:1910.14535 using synthetic data. We then verify that additional b-jets can be identified more accurately by increasing matching efficiency directly rather than the binary classification accuracy, using simulated $\text{t}\bar{\text{t}}\text{b}\bar{\text{b}}$ event data in the lepton+jets channel from pp collision at $\sqrt{s}$ = 13 TeV.

hep-ex↗

arXiv2021

Proceedings of the second MadAnalysis 5 workshop on LHC recasting in Korea

We document the activities performed during the second MadAnalysis 5 workshop on LHC recasting, that was organised in KIAS (Seoul, Korea) on February 12-20, 2020. We detail the implementation of 12 new ATLAS and CMS searches in the MadAnalysis 5 Public Analysis Database, and the associated validation procedures. Those searches probe the production of extra gauge and scalar/pseudoscalar bosons, supersymmetry, seesaw models and deviations from the Standard Model in four-top production.

hep-ph↗

arXiv2019

Identification of additional jets in the t-tbar b-bbar events using a deep neural network

In the top quark pair production in association with the Higgs boson decaying to a b quark pair t-tbar H (b-bbar), the final state has an irreducible nonresonant background from the production of a top quark pair in association with a b quark pair t-tbar b-bbar. Therefore, understanding of the t-tbar b-bbar process precisely in particular differential cross-section as functions of the properties of the additional b jets not from the top quark decay is essential for improving the sensitivity of a search for the t-tbar H b-bbar process. The two additional b jets can be identified by using various approaches. In this paper, the performances are compared quantitatively in the lepton+jets decay channel in terms of the matching efficiency of assigning two additional b jets as a figure of merit. This study provides valuable information towards the precise measurement of differential cross-sections as a function of properties of the additional b jets in the t-tbar b-bbar events. We showed that a matching efficiency of around 40% could be achieved using a deep neural network method. In the events with at least 4 b jets, this performance is 8% better than that achieved using minimum Delta R(b,bbar) method. This is consistent with the boosted decision tree method within its statistical uncertainty.

hep-ex↗