AI ethics evidence index Search
Source-linked daily record
Evidence record 19475 · automatically gathered

I-SDPO: Instance-Level Adaptive Self-Distillation Policy Optimization

Group Relative Policy Optimization (GRPO) learns from reward differences within a rollout group, but receives no useful relative signal when every sampled response is incorrect. Privileged self-distillation can fill this gap with dense token supervision, yet applying it throughout training creates a different failure mode: the teacher is a biased, low-variance surrogate for the reward objective, so persistent imitation can oppose reward-improving updates after the policy becomes capable of produ

Record details

Published: 13 August 2026
Source: arXiv cs.LG
Category: Research
Topics: Bias & fairness · Regulation
Retrieved: 14 August 2026

source-onlyevidence status

These records share source-supplied organisations, an exact publisher byline, automatic topics or regions. The reason is shown on every link; related does not mean supporting, agreeing with or verifying this record.

Shared topic: Bias & fairness · Shared topic: Regulation

Variable Selection in the Context of AI Fairness

arXiv cs.CY · 13 August 2026

Shared topic: Bias & fairness · Shared topic: Regulation

Language Models as a Challenge for Business Ethics – A Partially Open-Source Approach

Science and Engineering Ethics · 14 August 2026

Shared topic: Bias & fairness · Shared topic: Regulation

IADD-TR: Intervention-Aware Dynamics Decoupling with Targeted Regularization for Model-Based Reinforcement Learning

arXiv cs.LG · 11 August 2026

Shared topic: Bias & fairness · Shared topic: Regulation

Manipulation-Proof Oblivious Audits against Deceptive Model Providers

arXiv · 5 August 2026

Shared topic: Bias & fairness · Shared topic: Regulation

Variable Selection in the Context of AI Fairness

arXiv fairness query · 4 August 2026

Shared topic: Bias & fairness · Shared topic: Regulation

Progressive in Principle, Centrist in Practice: LLM Political Bias Is Instrument-Dependent

arXiv cs.CY · 4 August 2026

How to cite this record

ethics.ai (13 August 2026), “I-SDPO: Instance-Level Adaptive Self-Distillation Policy Optimization,” evidence record 19475, https://ethics.ai/record/19475 (originally published by arXiv cs.LG).

JSON

Use and limitations

This page is a stable index and citation surface for a source record. ethics.ai did not author the underlying report and has not independently verified every claim. Automatic topics may be imperfect. For consequential use, quote and cite the original publisher.