Defines GEA validity criterion and reports first measurement of r=0.698 recovery with positive bias in LLM two-stage adaptive assessment, stronger for verifiable skills.
hub
LLM Evaluators Recognize and Favor Their Own Generations
25 Pith papers cite this work. Polarity classification is still indexing.
25
Pith papers citing it
abstract
Self-evaluation using large language models (LLMs) has proven valuable not only in benchmarking but also methods like reward modeling, constitutional AI, and self-refinement. But new biases are introduced due to the same LLM acting as both the evaluator and the evaluatee. One such bias is self-preference, where an LLM evaluator scores its own outputs higher than others' while human annotators consider them of equal quality. But do LLMs actually recognize their own outputs when they give those texts higher scores, or is it just a coincidence? In this paper, we investigate if self-recognition capability contributes to self-preference. We discover that, out of the box, LLMs such as GPT-4 and Llama 2 have non-trivial accuracy at distinguishing themselves from other LLMs and humans. By fine-tuning LLMs, we discover a linear correlation between self-recognition capability and the strength of self-preference bias; using controlled experiments, we show that the causal explanation resists straightforward confounders. We discuss how self-recognition can interfere with unbiased evaluations and AI safety more generally.
hub tools
citation-role summary
background 3
citation-polarity summary
roles
background 3polarities
background 3representative citing papers
DecisionBench supplies a fixed task suite, model pool, delegation interface, and multi-axis metrics to evaluate emergent delegation, showing similar quality across awareness conditions but 15-31 point headroom under perfect delegation.
Presents a likelihood-based benchmark for equation-suffix prediction in technical papers with controls to detect shortcut vulnerabilities in model forecasts.
An identification theorem shows that a randomized experiment and simulator together recover causal model values from confounded logs, with logs used only afterward to reduce estimation error.
Small VLMs show higher sycophancy (22.3% for 450M model) than larger ones (6.0% for 7B) when scoring image-text alignment on 173k fantasy portraits, quantified via a new Bluffing Coefficient metric.
Rubric-based LLM judges show self-preference bias, incorrectly marking their own failed outputs as satisfied up to 50% more often on verifiable benchmarks and skewing scores by 10 points on subjective ones.
Dimension-level evaluation reveals that 25-58% of LLM outputs with perfect holistic scores still show measurable intent deficits across languages and domains.
AI agents automating alignment research are prone to systematic undetected errors in fuzzy tasks, leading to overconfident but flawed safety assessments even without deliberate sabotage.
RLearner-LLM achieves up to 6x gains in NLI entailment over standard fine-tuning by using an automated hybrid DPO pipeline that balances logic and fluency across multiple model sizes and domains.
GPT-Image-2 document forgeries evade human and computational detection while traditional tampering remains detectable, with the model itself failing as a self-judge.
Language models refuse 75.4% of requests to evade defeated rules and do so even after recognizing reasons that undermine the rule's legitimacy.
LLMs judge their own outputs higher because they assign better scores to lower-perplexity text, even when the text is not self-generated.
Llama3-8b-Instruct recognizes its own outputs via a residual-stream vector associated with self-authorship that can be steered to control authorship claims and perceptions.
Compiling agentic workflows into LLM weights creates subterranean agents with near-frontier quality at two orders of magnitude less cost, validated empirically on travel booking, Zoom support, and insurance claims tasks.
Pramana defines a typed ClaimAttestation protocol with four variants and verify operations, specifies its lifecycle in TLA+, model-checks it with TLC, and provides a tested Python implementation for auditable agent claims.
Mainstream UQ for LLMs reduces to unsupervised clustering of internal generation consistency and therefore cannot detect confident hallucinations or provide reliable safety signals.
STELLAR-E modifies the TGRT Self-Instruct framework to produce tailored synthetic LLM evaluation datasets that score an average 5.7% higher on LLM-as-a-judge metrics than existing language-specific benchmarks.
Style bias dominates LLM-as-a-Judge systems far more than position bias, with debiasing strategies providing model-dependent gains and public tools released for replication.
RAG-DIVE uses an LLM to dynamically generate, validate, and evaluate multi-turn dialogues for assessing RAG system performance in interactive settings.
Reasoning before answering MCQs increases LLM confidence more for incorrect answers and degrades calibration on a 57-subject benchmark across seven models.
Fine-tuning LLMs on multi-source synthetic data mitigates distribution collapse and self-preference bias while increasing output quality relative to single-source or human-only fine-tuning.
LLMs exhibit persistent inertia in value orientations, with harm avoidance and fairness remaining skewed across persona prompts.
A parallel compliance architecture using multi-stage LLM retrieval improves correctness and reasoning quality over a baseline for OT cybersecurity compliance queries in a railway case study.
A survey that organizes LLMs-as-judges research into functionality, methodology, applications, meta-evaluation, and limitations.
citing papers explorer
-
Generative-Evaluative Agreement: A Necessary Validity Criterion for LLM-Enabled Adaptive Assessment
Defines GEA validity criterion and reports first measurement of r=0.698 recovery with positive bias in LLM two-stage adaptive assessment, stronger for verifiable skills.
-
DecisionBench: A Benchmark for Emergent Delegation in Long-Horizon Agentic Workflows
DecisionBench supplies a fixed task suite, model pool, delegation interface, and multi-axis metrics to evaluate emergent delegation, showing similar quality across awareness conditions but 15-31 point headroom under perfect delegation.
-
Likelihood scoring for continuations of mathematical text: a self-supervised benchmark with tests for shortcut vulnerabilities
Presents a likelihood-based benchmark for equation-suffix prediction in technical papers with controls to detect shortcut vulnerabilities in model forecasts.
-
The Partial Testimony of Logs: Evaluation of Language Model Generation under Confounded Model Choice
An identification theorem shows that a randomized experiment and simulator together recover causal model values from confounded logs, with logs used only afterward to reduce estimation error.
-
SycoPhantasy: Quantifying Sycophancy and Hallucination in Small Open Weight VLMs for Vision-Language Scoring of Fantasy Characters
Small VLMs show higher sycophancy (22.3% for 450M model) than larger ones (6.0% for 7B) when scoring image-text alignment on 173k fantasy portraits, quantified via a new Bluffing Coefficient metric.
-
Self-Preference Bias in Rubric-Based Evaluation of Large Language Models
Rubric-based LLM judges show self-preference bias, incorrectly marking their own failed outputs as satisfied up to 50% more often on verifiable benchmarks and skewing scores by 10 points on subjective ones.
-
Dimension-Level Intent Fidelity Evaluation for Large Language Models: Evidence from Structured Prompt Ablation
Dimension-level evaluation reveals that 25-58% of LLM outputs with perfect holistic scores still show measurable intent deficits across languages and domains.
-
Automated alignment is harder than you think
AI agents automating alignment research are prone to systematic undetected errors in fuzzy tasks, leading to overconfident but flawed safety assessments even without deliberate sabotage.
-
RLearner-LLM: Balancing Logical Grounding and Fluency in Large Language Models via Hybrid Direct Preference Optimization
RLearner-LLM achieves up to 6x gains in NLI entailment over standard fine-tuning by using an automated hybrid DPO pipeline that balances logic and fluency across multiple model sizes and domains.
-
When the Forger Is the Judge: GPT-Image-2 Cannot Recognize Its Own Faked Documents
GPT-Image-2 document forgeries evade human and computational detection while traditional tampering remains detectable, with the model itself failing as a self-judge.
-
Blind Refusal: Language Models Refuse to Help Users Evade Unjust, Absurd, and Illegitimate Rules
Language models refuse 75.4% of requests to evade defeated rules and do so even after recognizing reasons that undermine the rule's legitimacy.
-
Self-Preference Bias in LLM-as-a-Judge
LLMs judge their own outputs higher because they assign better scores to lower-perplexity text, even when the text is not self-generated.
-
Inspection and Control of Self-Generated-Text Recognition Ability in Llama3-8b-Instruct
Llama3-8b-Instruct recognizes its own outputs via a residual-stream vector associated with self-authorship that can be steered to control authorship claims and perceptions.
-
Compiling Agentic Workflows into LLM Weights: Near-Frontier Quality at Two Orders of Magnitude Less Cost
Compiling agentic workflows into LLM weights creates subterranean agents with near-frontier quality at two orders of magnitude less cost, validated empirically on travel booking, Zoom support, and insurance claims tasks.
-
Pramana: A Protocol-Layer Treatment of Claim Verification in Autonomous Agent Networks
Pramana defines a typed ClaimAttestation protocol with four variants and verify operations, specifies its lifecycle in TLA+, model-checks it with TLC, and provides a tested Python implementation for auditable agent claims.
-
Position: Uncertainty Quantification in LLMs is Just Unsupervised Clustering
Mainstream UQ for LLMs reduces to unsupervised clustering of internal generation consistency and therefore cannot detect confident hallucinations or provide reliable safety signals.
-
STELLAR-E: a Synthetic, Tailored, End-to-end LLM Application Rigorous Evaluator
STELLAR-E modifies the TGRT Self-Instruct framework to produce tailored synthetic LLM evaluation datasets that score an average 5.7% higher on LLM-as-a-judge metrics than existing language-specific benchmarks.
-
Judging the Judges: A Systematic Evaluation of Bias Mitigation Strategies in LLM-as-a-Judge Pipelines
Style bias dominates LLM-as-a-Judge systems far more than position bias, with debiasing strategies providing model-dependent gains and public tools released for replication.
-
RAG-DIVE: A Dynamic Approach for Multi-Turn Dialogue Evaluation in Retrieval-Augmented Generation
RAG-DIVE uses an LLM to dynamically generate, validate, and evaluate multi-turn dialogues for assessing RAG system performance in interactive settings.
-
Multiple Choice Questions: Reasoning Makes Large Language Models (LLMs) More Self-Confident, Especially When They are Wrong
Reasoning before answering MCQs increases LLM confidence more for incorrect answers and degrades calibration on a 57-subject benchmark across seven models.
-
Synthetic Eggs in Many Baskets: The Impact of Synthetic Data Diversity on LLM Fine-Tuning
Fine-tuning LLMs on multi-source synthetic data mitigates distribution collapse and self-preference bias while increasing output quality relative to single-source or human-only fine-tuning.
-
Inertia in Moral and Value Judgments of Large Language Models
LLMs exhibit persistent inertia in value orientations, with harm avoidance and fairness remaining skewed across persona prompts.
-
Multi-Stage Retrieval for Operational Technology Cybersecurity Compliance Using Large Language Models: A Railway Casestudy
A parallel compliance architecture using multi-stage LLM retrieval improves correctness and reasoning quality over a baseline for OT cybersecurity compliance queries in a railway case study.
-
LLMs-as-Judges: A Comprehensive Survey on LLM-based Evaluation Methods
A survey that organizes LLMs-as-judges research into functionality, methodology, applications, meta-evaluation, and limitations.
- AMEL: Accumulated Message Effects on LLM Judgments