REVIEW 31 cited by
Aligning Large Language Models with Human: A Survey
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
Large Language Models (LLMs) trained on extensive textual corpora have emerged as leading solutions for a broad array of Natural Language Processing (NLP) tasks. Despite their notable performance, these models are prone to certain limitations such as misunderstanding human instructions, generating potentially biased content, or factually incorrect (hallucinated) information. Hence, aligning LLMs with human expectations has become an active area of interest within the research community. This survey presents a comprehensive overview of these alignment technologies, including the following aspects. (1) Data collection: the methods for effectively collecting high-quality instructions for LLM alignment, including the use of NLP benchmarks, human annotations, and leveraging strong LLMs. (2) Training methodologies: a detailed review of the prevailing training methods employed for LLM alignment. Our exploration encompasses Supervised Fine-tuning, both Online and Offline human preference training, along with parameter-efficient training mechanisms. (3) Model Evaluation: the methods for evaluating the effectiveness of these human-aligned LLMs, presenting a multifaceted approach towards their assessment. In conclusion, we collate and distill our findings, shedding light on several promising future research avenues in the field. This survey, therefore, serves as a valuable resource for anyone invested in understanding and advancing the alignment of LLMs to better suit human-oriented tasks and expectations. An associated GitHub link collecting the latest papers is available at https://github.com/GaryYufei/AlignLLMHumanSurvey.
Forward citations
Cited by 31 Pith papers
-
CrossPipe: Towards Optimal Pipeline Schedules for Cross-Datacenter Training
CrossPipe generates latency- and bandwidth-aware pipeline schedules that cut emulated cross-datacenter LLM training time by up to 33.6%.
-
Don't Mix Rewards, Mix Policies: Policy Decomposition and Optimization for Multi-Reward RL
PRISM trains one positive policy per reward plus one global negative policy and merges their token logits, improving multi-reward RL for LLMs with inference-time controllability.
-
Probably Correct Optimal Stable Matching under Two-Sided Uncertainty
Elimination algorithms identify the optimal stable matching with high probability under two-sided uncertainty by exploiting partial preferences and pervasive stable matchings, yielding sample-complexity and regret bou...
-
Facial-Expression-Aware Prompting for Empathetic LLM Tutoring
Feeding LLM tutors a text description or AUM-selected frame of a student's facial expression improves rated empathetic responsiveness across three backbones.
-
GUI-Libra: Training Native GUI Agents to Reason and Act with Action-aware Supervision and Partially Verifiable RL
Action-weighted SFT plus KL-regularized GRPO on a curated 81K reasoning dataset lifts open-source GUI agents by 11–22 points on online task-completion benchmarks.
-
SharedRep-RLHF: A Shared Representation Approach to RLHF with Diverse Preferences
SharedRep-RLHF learns a shared preference representation across groups to improve worst-case reward estimates for minority annotators, but the theoretical guarantees are undermined by proof errors.
-
Exploring the Impact of Instruction-Tuning on LLM's Susceptibility to Misinformation
Instruction tuning shifts LLM susceptibility to misinformation from assistant-provided text to user-provided text, making user-role misinformation more likely to override the model's correct parametric knowledge.
-
Beyond Reactive Safety: Risk-Aware LLM Alignment via Long-Horizon Simulation
LLM-based long-horizon event simulation, used as a reward signal, is claimed to improve safety alignment and indirect-harm detection, but evaluation confounds simulation with the capability of the external projector model.
-
Structured Moral Reasoning in Language Models: A Value-Grounded Evaluation Framework
Structured moral prompts, especially first-principles reasoning, improve LLM moral classification accuracy across 12 open models and four benchmarks, and reasoning distillation transfers these gains to a 3B model.
-
From Outcomes to Processes: Guiding PRM Learning from ORM for Inference-Time Alignment
SP-PRM trains process reward models from outcome reward models via score and preference consistency, improving reward-guided search on dialogue, summarization, and reasoning tasks.
-
Tag-Evol: Achieving Efficient Instruction Evolving via Tag Injection
Tag-Evol generates harder, more diverse instruction data by injecting sampled knowledge tags into seed instructions, improving downstream SFT accuracy across math, code, and general benchmarks.
-
Self-Consistency of the Internal Reward Models Improves Self-Rewarding Language Models
Forcing a self-rewarding LLM's generative judge and implicit DPO reward to agree on preference labels substantially improves its AlpacaEval 2.0 alignment score.
-
JBShield: Defending Large Language Models from Jailbreak Attacks through Activated Concept Analysis and Manipulation
JBShield detects jailbreaks by checking whether a prompt activates both a toxic concept and a jailbreak concept inside an LLM, then steers those concepts to produce a safe refusal.
-
SOS-LoRA: Static Orthogonal-Subspace Low-Rank Adaptation with Fixed Multi-Scale Scaling
A LoRA update split into several fixed, differently-scaled low-rank experts with orthogonal input directions improves fine-tuning accuracy at the same parameter count.
-
Toward Preference-aligned Large Language Models via Residual-based Model Steering
Preference signals in LLM residual streams can be distilled into inference-time steering vectors that improve math and code benchmarks using only 100 preference pairs.
-
Path to Intelligence: Measuring Similarity between Human Brain and Large Language Model Beyond Language Task
LLM hidden states can be linearly projected onto human intracranial EEG recorded during a text-translated sensory-motor anticipation task, with moderate CKA alignment and similar response-time distributions.
-
An Uncertainty-Driven Adaptive Self-Alignment Framework for Large Language Models
A training pipeline that scores a model's own responses for semantic, factual, and safety uncertainty, builds preference pairs from those scores, and trains in three difficulty stages improves reported alignment score...
-
Token-level Accept or Reject: A Micro Alignment Approach for Large Language Models
MARA aligns LLMs with human preferences by training a 4M-parameter MLP to accept or reject candidate tokens, avoiding full-model fine-tuning, with measured gains based on the same reward models used in training.
-
One Model Transfer to All: On Robust Jailbreak Prompts Generation against LLMs
ArrAttack fine-tunes a judge on the SmoothLLM defense, uses it to filter rewriting-attack data, and trains a generator that produces jailbreak prompts transferring across defenses.
-
Shallow Preference Signals: Large Language Model Aligns Even Better with Truncated Data?
Preference signals in LLM alignment are concentrated in early response tokens, so models trained on data truncated to the first half perform as well as or better than those trained on full responses.
-
Data2Concept2Text: An Explainable Multilingual Framework for Data Analysis Narration
A six-stage Prolog/CLP tree-rewriting pipeline converts concept trees into multilingual natural language sentences with explicit rule traces, demonstrated on data narration and an ICLP call-for-papers example.
-
Salamandra Technical Report
Salamandra is an open, from-scratch multilingual LLM family with 2B, 7B, and 40B checkpoints, instruction-tuned variants, a vision proof-of-concept, and detailed evaluations across Iberian and European languages.
-
LLM Alignment as Retriever Optimization: An Information Retrieval Perspective
LARPO, an iterative preference optimization method that adapts information retrieval techniques such as listwise ranking losses, hard negatives, and candidate lists, is claimed to substantially improve LLM alignment o...
-
FActBench: A Benchmark for Fine-grained Automatic Evaluation of LLM-Generated Text in the Medical Domain
Unanimous voting between NLI and chain-of-thought fact-checking yields scores closest to medical expert judgments on three of four tasks in the new FActBench benchmark.
-
A Comprehensive Evaluation framework of Alignment Techniques for LLMs
The paper proposes a multi-dimensional framework to evaluate and compare LLM alignment techniques.
-
The Fair Game: Auditing & Debiasing AI Algorithms Over Time
Proposes 'Fair Game', a reinforcement-learning loop in which an auditor's bias criteria, updatable over time, steer a debiasing agent that adapts an ML model's predictions.
-
Reliable Annotations with Less Effort: Evaluating LLM-Human Collaboration in Search Clarifications
LLMs alone annotate search clarifications unreliably; adding confidence-based selective human review cuts effort 24-45% in simulation, but the evaluation is partly built from the ground truth it predicts.
-
Crowd-SFT: Crowdsourcing for LLM Alignment
A competitive multi-group fine-tuning framework with point rewards correlated to Shapley values reduced simulated model distance by up to 55% and tracked user contributions reasonably in vector-space experiments.
-
Large Language Models for EEG: A Comprehensive Survey and Taxonomy
A taxonomy and review of studies applying large language models to EEG signals, organized into four domains and three adaptation strategies.
-
The Science of Evaluating Foundation Models
A survey-and-checklist proposal that organizes LLM evaluation into an ABCD framework (Algorithm, Big Data, Computation, Domain Expertise) for context-aware, documented assessment.
-
Linguistics and Human Brain: A Perspective of Computational Neuroscience
A narrative review arguing that computational neuroscience, powered by LLM-based model–brain alignment, serves as the bridge between linguistic theory and neural data.
Discussion (0). Continue with ORCID to comment.