FORT synthesizes shortcut-resistant search tasks by controlling four identified shortcut risks across entity selection, graph construction, question formulation, and refinement, producing training data that yields agents with longer search trajectories and top performance among open-source models on
arXiv preprint arXiv:2602.10604 , year=
13 Pith papers cite this work. Polarity classification is still indexing.
13
Pith papers citing it
citation-role summary
background 2
baseline 1
method 1
citation-polarity summary
years
2026 13representative citing papers
Introduces KINA benchmark with 899 items over 261 disciplines, formal (1-1/e) coverage guarantee and bonus-on-bar tournament theorem, plus evaluations of 42 models with top score 53.17%.
ProofRank benchmark shows substantial differences in LLM proof quality not captured by correctness, with trade-offs between quality metrics and accuracy.
OPHSD uses harness-augmented models as teachers to distill reasoning capabilities into base LLMs, yielding strong standalone performance on classification and math tasks.
The cancellation hypothesis shows how rollout-level rewards produce token-level credit assignment in critic-free RL through cancellation of opposing signals on shared tokens, with empirical support and batching interventions that enhance performance.
Self-play between LLMs for problem authoring and solving, scored via Rasch modeling, shows that authoring and solving skills are partially decoupled and that the benchmark difficulty evolves with new models.
MIR improves validation loss in repeated-data pretraining and SoftQ fits data-constrained scaling experiments better than additive laws, equating MIR gains to roughly 1.3 times more unique data.
DFlare replaces DFlash's shared fused representation with per-draft-layer attention to distinct target-layer combinations, enabling deeper drafts and 2.4M training samples for 5-11% higher speedups than DFlash on Qwen3 and GPT-OSS models.
Proposes equivariant optimizer updates matched to layer symmetries for embeddings, SwiGLU MLPs, and MoE routers, with reported gains in validation loss and training stability on several language model architectures.
PrfaaS enables practical cross-datacenter prefill-decode disaggregation for hybrid-attention models via selective offloading, bandwidth-aware scheduling, and cache-aware placement, yielding 54% higher throughput and 64% lower P90 TTFT than homogeneous baselines in a 1T-parameter case study.
Attention Editing converts pre-trained LLMs to new attention architectures through layer-wise teacher-forced optimization and model-level distillation, preserving performance with efficiency gains.
CODA uses rollout-based difficulty signals to drive two gates that penalize verbosity on easy instances and promote deliberation on hard ones, cutting token use over 60% on simple tasks while maintaining accuracy.
InternVideo3 introduces Multimodal Contextual Reasoning and M^2LA attention to enable closed-loop evidence accumulation in long-video understanding and agentic tool use, reporting strong benchmark results.
citing papers explorer
-
FORT-Searcher: Synthesizing Shortcut-Resistant Search Tasks for Training Deep Search Agents
FORT synthesizes shortcut-resistant search tasks by controlling four identified shortcut risks across entity selection, graph construction, question formulation, and refinement, producing training data that yields agents with longer search trajectories and top performance among open-source models on
-
Not All Proofs Are Equal: Evaluating LLM Proof Quality Beyond Correctness
ProofRank benchmark shows substantial differences in LLM proof quality not captured by correctness, with trade-offs between quality metrics and accuracy.
-
Training with Harnesses: On-Policy Harness Self-Distillation for Complex Reasoning
OPHSD uses harness-augmented models as teachers to distill reasoning capabilities into base LLMs, yielding strong standalone performance on classification and math tasks.
-
MathDuels: Evaluating LLMs as Problem Posers and Solvers
Self-play between LLMs for problem authoring and solving, scored via Rasch modeling, shows that authoring and solving skills are partially decoupled and that the benchmark difficulty evolves with new models.
-
DFlare: Scaling Up Draft Capacity for Block Diffusion Speculative Decoding
DFlare replaces DFlash's shared fused representation with per-draft-layer attention to distinct target-layer combinations, enabling deeper drafts and 2.4M training samples for 5-11% higher speedups than DFlash on Qwen3 and GPT-OSS models.
-
CODA: Difficulty-Aware Compute Allocation for Adaptive Reasoning
CODA uses rollout-based difficulty signals to drive two gates that penalize verbosity on easy instances and promote deliberation on hard ones, cutting token use over 60% on simple tasks while maintaining accuracy.