Instruction Tuning

Follow this topic Ask about this topic

Instruction tuning studies how to turn pretrained models into systems that respond usefully to instructions, often through supervised fine-tuning on instruction-response pairs and sometimes preference optimization or reinforcement learning. Research develops ways to select, balance, and generate training data, including methods for multilingual, domain-specific, and multimodal tasks. Evaluations use instruction-following tests such as IFEval alongside reasoning, coding, dialogue, and specialized-domain benchmarks. A central challenge is that gains can depend on task and format: tuning may improve instruction compliance while harming other capabilities, and reliable generalization beyond training examples remains difficult.

Latest papers

81 papers13 papers in the last four weeks, up 86% on the four weeks before. 0.1% of all new papers.

The Geometry of Emotion Representations Across Scale, Depth, and Instruction Tuning
Oct 26, 2026Adam HollowellEmotion Representation in Language Models Representation Geometry in Language Models
University of North Carolina at Chapel Hill
DocTalk: Enhancing Medical Knowledge Retrieval via Fine-tuned Large Language Models with Chain-of-Thought Reasoning for Robust Clinical Decision Support
Oct 26, 2026Asad Gohar, Md Ariful Islam Mozumder, Moon-Il Joo +1Clinical QA EHR QA
Institute of Digital Anti-Aging HealthCare, Inje University, Gimhae 50834, Republic of Korea
"High-Quality" Data for Early-Stage Instruction Tuning: An Empirical Study of Data Quality Metrics
Oct 25, 2026Vladimir Necula, David Gvadzabia, Sofia SerranoInstruction-Tuning Data Selection Instruction Tuning
University of Cambridge Department of Physics · Lafayette College Department of Computer Science
The Language of Instruction: Genre and Generative AI
Oct 23, 2026James E. DobsonLLM Prompting Instruction Tuning
Department of English and Creative Writing Dartmouth College
Task-Preserving Multilingual Instruction Tuning for Open-Ended QA in WMT 2026 MIST
Oct 21, 2026Aleix Sant, Jordi Luque, Carlos EscolanoMultilingual QA Fine-Tuning Language Models
Scientific Research, Telefónica Innovación Digital · Universitat Politècnica de Catalunya
The Instruction Tax on Factual Recall: Base Models Can Beat Their Instruction-Tuned Counterparts
Oct 20, 2026Christophe Khalil, Aya MouradFactual Knowledge in Language Models Instruction Tuning
Sorbonne Université, Paris, France
LongForge: Measuring and Improving Long-Context Instruction Data via Evidence-Assertion Reasoning Graphs
Oct 16, 2026Ziyang Chen, Junlong Jia, Xing Wu +2Synthetic Data Generation for Language Models Instruction Tuning
AutoSFT: An Agentic System for Automated Data Research in Supervised Fine-Tuning
Oct 15, 2026Bingxiang He, Pengqi Sun, Shihao Peng +18Supervised Fine-Tuning Instruction-Tuning Data Selection
Tsinghua University · Beijing University of Posts and Telecommunications · ModelBest Inc. +2
Multi-SFT: Permissively-Licensed Multilingual Instruction-Following
Oct 15, 2026Daniela Baiamonte, Kamyar Zeinalipour, Elena Fano +2Instruction Tuning Multilingual Language Models
Yukai! · University of Siena
The Format is the Filter: Scoring Format Fixes What an LLM Judge Can Express
Oct 12, 2026Lovedeep Gondara, Jekaterina Novikova, Sagarika Suresh +2LLM-as-a-Judge Bias in LLM-as-a-Judge
The Vanguard Group, Inc.
PitchIn: Natural-Language Vocal Coaching from Unstructured Community Feedback
Oct 12, 2026Hayeon Bang, Hounsu Kim, Joonyong Park +2Audio-Language Models Audio Understanding
KAIST, Republic of Korea · The University of Tokyo, Japan · ETRI, Republic of Korea
Enhancing Biomedical Named Entity Recognition via Multiple Programming Languages Instruction Tuning and Ensemble Method
Oct 2, 2026Songtao Li, Yijia Zhang, Jianyuan Yuan +3Named Entity Recognition LLM Fine-Tuning
School of Information Science and Technology, Dalian Maritime University, Dalian, China · School of Automation, Beijing Institute of Technology, Beijing, China · School of Computer Science and Engineering, Northeastern University, Shenyang, China +1
Coverage Before Control: Route-Instruction Grounding and Steering for Controllable Retrosynthesis
Sep 30, 2026Xuemin Chen, Xiaozhuang Song, Xinjian Zhao +2Instruction Following Instruction Tuning
The Chinese University of Hong Kong, Shenzhen
A helps B while B hurts A: directed transfer in instruction-tuning mixture
Sep 30, 2026Nima H. Siboni, Vahid RostamiLLM Fine-Tuning Instruction-Tuning Data Selection
Juna.ai, Kastanienallee 32, 10435 Berlin, Germany · Computational Systems Neuroscience, Institute of Zoology, University of Cologne, Germany
CompOrca: Corpus-Scale Compliance Labelling of Instruction-Tuning Data
Sep 29, 2026Philipp E. Glass, Alina MironLLM Auditing Instruction Tuning
Department of Computer Science, Brunel University of London
ScopeIF: Improving Scope-Aware Precise Instruction-Following in Large Language Models via Graded Reward Modeling
Sep 26, 2026Bosi Wen, Yilin Niu, Xiaoying Ning +3Instruction Following Instruction Tuning
The Conversational Artificial Intelligence (CoAI) Group, Tsinghua University · Zhipu AI
TemporalGraphLLM: Temporal Graph Neural Networks with Large Language Models for Dynamic Text-Attributed Graphs
Sep 25, 2026Moran Beladev, Or Eitan, Gilad Katz +1Text-Attributed Graph Learning LLMs for Graph Learning
Department of Information Systems Engineering Ben-Gurion University of the Negev, Israel
Rewired or Gated? How Instruction Tuning Shapes Knowledge-Conflict Circuits in LLMs
Sep 22, 2026Shubham Santosh Pandere, Gautam Ranka, Ritika Varshney +3Knowledge Conflicts in Language Models LLM Interpretability
IvLabs, VNIT
Clustering-Based Balanced Sampling and Allocation with Data Parallelism for High-Performance Fine-Tuning
Sep 11, 2026Hyunjin Kim, Youngeun Nam, Jaemin Han +2LLM Fine-Tuning Instruction-Tuning Data Selection
KAIST
Stabilizing Instruction Supervision for Instruct-TTS via Controllable Diversification and Drift Filtering
Sep 8, 2026Yizhong Geng, Kecan Mao, Qifei Li +6TTS Evaluation Training Data Curation
Beijing University of Posts and Telecommunications, China · Li Auto, China
EuroAlpaca: Task-Preserving Localisation of Instruction Data for European Languages
Sep 4, 2026Aleix Sant, Jordi Luque, Carlos EscolanoInstruction Tuning Machine Translation
Scientific Research, Telefónica Innovación Digital · Universitat Politècnica de Catalunya Barcelona, Spain
How Output Format Confounds Data Quality and Capability in Instruction Tuning
Sep 2, 2026Chengguang Gan, Hanjun Wei, Yunhao Liang +3LLM Evaluation Instruction Tuning
1Independent Researcher · University of Chinese Academy of Sciences · 3Pusan National University +2
From Specialization to Generalization: Instruction-tuned LLMs for Robust Harmful Content Mitigation
Aug 26, 2026Lukas Edman, Daryna Dementieva, Alexander FraserContent Moderation Hate Speech Detection
School of Computation, Information and Technology, TU Munich · Munich Center for Machine Learning · Munich Data Science Institute
Industrial-Instruction: An End-to-End Framework for Building Instruction-Tuning and Benchmark Datasets from Industrial Technical Reports
Aug 24, 2026Parsa Bakhtiari, Hassan Bashiri, Alireza Khalilipour +2Synthetic Benchmark Generation Training Data Extraction
Department of Computer Engineering, Hamedan University of Technology, Hamedan, Iran. · Department of Computer Science, University of Antwerp, Antwerp, Belgium; Flanders Make Strategic Research Center, Antwerp, Belgium.
Are You Sure You're Sure? On the Impact of Instruction Tuning on Confidence and Lexical Diversity
Aug 13, 2026Irina Proskurina, Mayank Kumar, Oyindolapo O. KomolafeConfidence Estimation in Language Models Question Answering
Cohere Labs Community · Laboratoire Hubert Curien, UMR CNRS 5516, Saint-Étienne, France · School of Computer Science Engineering and Technology (SCSET), Bennett University, Greater Noida, India +1
Agentic Instruction Data Selection: Let DataMaster Interpret Your Intent
Aug 11, 2026Fanqi Zhou, Qiaosheng Chen, Zixian Huang +1Training Data Curation Instruction-Tuning Data Selection
State Key Laboratory of Novel Software Technology, Nanjing University · Shanghai Artificial Intelligence Laboratory
CreativeInstruct: Scalably Teaching LLMs to Balance Quality, Creativity, and Diversity
Aug 7, 2026Ananya Sahu, Mohit Bansal, Elias Stengel-EskinInstruction Tuning Diverse Text Generation
Columbia · UNC Chapel Hill · University of Texas at Austin
HSS-Synth: Humanities and Social Sciences Data Synthesis for LLMs
Jul 29, 2026Ru Peng, Tianyu Zhao, Xijun Gu +9LLM Fine-Tuning Instruction Tuning
Zhejiang University · Inclusion AI, Ant Group · Ant Group +2
Instruction-Tuned Models Locally Reuse Human Syntax More Than Humans Do
Jul 28, 2026Zandi EberstadtLanguage Modeling Instruction Tuning
Department of Computer Science University of Oxford
A Unified Moral-Value Dataset for Instruction Tuning
Jul 23, 2026Zhaohui Zeng, Florian MaiLLM Alignment Moral Reasoning in Language Models
1RWTH Aachen · University of Bonn · 3Lamarr Institute for Machine Learning and Artificial Intelligence

Oct 26, 2026J/K move · Enter open · S save

The Geometry of Emotion Representations Across Scale, Depth, and Instruction Tuning

Adam Hollowell

University of North Carolina at Chapel Hill

Emotion Representation in Language Models Representation Geometry in Language Models Instruction Tuning

Open paper AskPDF arXiv

We measure the emotion geometry of six LLMs spanning Google's Gemma 2 family. Using an extraction pipeline established by Sofroniew et al. (2026) on Claude Sonnet 4.5 and replicated by Hollowell (2026) on Gemma 2 27B base, we compare emotion geometry across base and instruction-tuned variants of Gemma 2 2B, 9B, and 27B. First, all six model cells exhibit the core emotion geometry. A linear map fit between two cells places held-out emotions nearest their own vectors in the target model. Across all fifteen model pairs, fit in one direction each, this failed exactly once, and the median rank was 1 in every pair. Linear maps refit on cleaned embedding-table vectors recover every held-out emotion at rank 1 in all thirty ordered pairs. Second, scale spreads the geometry's variance across more directions without weakening its tie to valence. The leading component's share of variance falls from 31.6 percent (2B) to 29.8 (9B) to 26.7 (27B) in the base models and from 31.9 (2B) to 29.2 (9B) to 27.3 (27B) in the tuned ones, while its correlation with human valence ratings stays between .71 and .78 in all six models. Third, we project each emotion vector onto the output vocabulary and ask at what depth the vector's own emotion word enters the top 10,000 of the 256,000 tokens. In the 2B and 9B models the median emotion word is there from layer 0, while in the 27B models it begins near chance, stays outside the top 10,000 for half the network, and enters at layer 23 in one step. Fourth, tuning leaves every registered analysis-layer measure within its floor. At every size, valence correlation rises and arousal correlation falls by amounts within their floors, the latter against prediction. PC2's variance share also falls, with a pooled decrease significant after correction. Mid-late layer stability exceeds its floor and traces to base 27B's final layer. Code, data, and audit reports are released.