arXiv:2608. 08164v1 Announce Type: cross Abstract: Knowledge Distillation is a widely adopted technique in the training and fine-tuning of large language models (LLMs) enabling transfer of structured information and functional behavior from a large teacher model to a smaller student model while significantly reducing computational costs.
The paper questions whether large language models (LLMs) truly introspect by critiquing recent studies that claim they can detect and report their internal states. It proposes two necessary conditions for genuine introspection: privileged access to internal representations and second‑order computation that distinguishes from first‑order task performance. Re‑examining two existing paradigms, the authors find that apparent introspective abilities can be explained by input‑based classifiers or generic anomaly detection, concluding that current evidence does not support metacognitive monitoring in LLMs.
Post-training large language models (LLMs) without real-world interaction feedback or human-labeled supervision remains challenging, particularly in specialized domains where expert annotations are costly to obtain. Recent annotation-free self-evolution methods address this by using the model's own outputs as supervision signals, constructing a teacher via additional context and aggregating predictions across multiple rollouts through majority voting to produce pseudo-labels.
Negative Self-Distillation (NSD) is a new framework for improving large language models by encouraging them to diverge from their own flawed reasoning rather than imitate privileged solutions. Unlike On-Policy Self-Distillation, which can suppress uncertainty and exploratory behavior, NSD generates a question‑specific negative condition (e.g., a careless reasoner) and uses a dynamic gating mechanism to target only reasoning‑critical tokens for penalization. This approach preserves foundational language capabilities while consistently outperforming OPSD and other label‑free self‑bootstrapping reinforcement learning baselines.
arXiv:2607. 02460v1 Announce Type: cross Abstract: Post-training large language models (LLMs) without real-world interaction feedback or human-labeled supervision remains challenging, particularly in specialized domains where expert annotations are costly to obtain.
Negative Self-Distillation (NSD) is a new framework for improving large language models by encouraging them to diverge from their own flawed reasoning rather than imitate privileged solutions. Unlike On-Policy Self-Distillation (OPSD), which can suppress uncertainty and exploratory behavior, NSD generates a question‑specific negative condition (e.g., a careless reasoner) and pushes the student’s distribution away from it. A dynamic gating mechanism isolates reasoning‑critical tokens so that only behavioral flaws are penalized, preserving linguistic capabilities, and empirical results show NSD consistently outperforms OPSD and other label‑free self‑bootstrapping RL baselines.
arXiv:2607. 23379v1 Announce Type: cross Abstract: Activation Oracles (AOs) are language models trained to answer natural-language questions about another model's internal activations.
The study investigates whether large language models (LLMs) can reliably detect when their own responses have been manipulated by adversarial prefill attacks. Across ten instruction‑tuned LLMs ranging from 3B to 70B parameters and four safety benchmarks, none consistently recognized compromised outputs, with models claiming intent on prefilled responses at an average of 25.3%. The research identifies that introspective signals mainly arise from safety reasoning and refusal, and that training to improve introspection can paradoxically increase attack success, underscoring the fragility of LLM self‑reporting in safety contexts.
arXiv:2607. 04428v1 Announce Type: cross Abstract: Diffusion large language models (dLLMs) generate text by iteratively denoising a masked sequence, offering a parallel alternative to autoregressive models, but eliciting strong reasoning through post-training remains difficult: supervised fine-tuning is off-policy and suffers from exposure bias, while reinforcement learning gives only sparse, sequence-level rewards and is hard to apply without tractable sequence likelihoods.
arXiv:2605.27186v2 Announce Type: replace Abstract: Large language models often solve tasks from a fully specified prompt but degrade when the same requirements unfold over multiple turns, known as t...
The paper introduces a method for identifying the dominant knowledge source behind large language model (LLM) outputs, distinguishing between faithfulness violations (misuse of provided context) and factuality violations (errors in internal knowledge). A simple linear probe trained on hidden representations can reliably classify this source, and the authors present AttriWiki, a self‑supervised pipeline that generates labeled training data by prompting models to recall withheld entities or read them from context. Probes trained on AttriWiki achieve high Macro‑F1 scores across several models and datasets, generalize zero‑shot to a benchmark, and show that attribution mismatches can increase error rates by up to 70%. "whyItMatters":"The study demonstrates that knowing the source of an LLM’s answer is crucial for effective mitigation of hallucinations, as attribution mismatches significantly raise error rates."
The paper investigates on‑policy self‑distillation (OPSD) as a method to enhance reasoning in language models, focusing on mathematical reasoning across models from 0.6B to 8B parameters. Through controlled experiments and token‑level analysis, the authors find that OPSD’s effectiveness depends on alignment between the teacher’s reasoning mode and the full teacher prefix, rather than on privileged semantics alone. They observe that OPSD only improves reasoning in limited compatibility regimes, while often causing length growth, degradation, or behavioral collapse, and that the teacher’s signal is unstable and not predictive of downstream performance.