PROPEL amortizes solver evaluation with a trained activation probe to optimize task generators toward a target solve rate, raising the share of learnable tasks from ~10% to ~20% in coding and SWE experiments.
Reasoning-finetuning repurposes latent representations in base models.arXiv:2507.12638
5 Pith papers cite this work. Polarity classification is still indexing.
5
Pith papers citing it
years
2026 5representative citing papers
Reward-guided optimization of SAE latents, gated by reward and confidence, improves multi-benchmark reasoning and is post-hoc associated with more verification and course correction.
CoRP consolidates reward-weighted perturbations into a single model via low-rank structure, improving base LLMs by 8.1 points on average while using one-tenth the budget of prior ensembles and one forward pass.
LLMs implement a second-order confidence architecture where the PANL activation encodes both error likelihood and the ability to correct it, beyond verbal confidence or log-probabilities.
A new pipeline uses interpretability to characterize concepts in preference data and shape rewards via feature or data interventions during LM post-training.
citing papers explorer
-
Breaking the Solver Bottleneck: Training Task Generators at the Learnable Frontier
PROPEL amortizes solver evaluation with a trained activation probe to optimize task generators toward a target solve rate, raising the share of learnable tasks from ~10% to ~20% in coding and SWE experiments.
-
Latent Reward Steering: An Adaptive Inference-Time Framework that Implicitly Promotes Cognitive Behaviors in Reasoning LLMs
Reward-guided optimization of SAE latents, gated by reward and confidence, improves multi-benchmark reasoning and is post-hoc associated with more verification and course correction.
-
Consolidating Rewarded Perturbations for LLM Post-Training
CoRP consolidates reward-weighted perturbations into a single model via low-rank structure, improving base LLMs by 8.1 points on average while using one-tenth the budget of prior ensembles and one forward pass.
-
How LLMs Detect and Correct Their Own Errors: The Role of Internal Confidence Signals
LLMs implement a second-order confidence architecture where the PANL activation encodes both error likelihood and the ability to correct it, beyond verbal confidence or log-probabilities.
-
Anatomy of Post-Training: Using Interpretability to Characterize Data and Shape the Learning Signal
A new pipeline uses interpretability to characterize concepts in preference data and shape rewards via feature or data interventions during LM post-training.