Verbal confidence in LLMs tracks future commit/abstain decisions more than answer correctness, while log-probabilities track correctness.
hub
org/abs/2601.01828
14 Pith papers cite this work. Polarity classification is still indexing.
14
Pith papers citing it
hub tools
citation-role summary
background 2
citation-polarity summary
years
2026 14roles
background 2representative citing papers
ContextEcho benchmark shows persona drift occurs across 23 frontier models in long agentic-coding sessions, is not reliably reset by compaction, and can be restored by single-shot anchors with mode-dependent effects.
CopT reverses CoT by eliciting a draft answer first then using continuous-embedding contrastive verification and on-policy thinking to reflect and correct, yielding up to 23% higher accuracy and 57% fewer tokens without training.
RouteHijack is a routing-aware jailbreak that identifies safety-critical experts via activation contrast and optimizes suffixes to suppress them, reaching 69.3% average attack success rate on seven MoE LLMs with strong transfer to variants and VLMs.
HydraHead hybridizes full and linear attention along the head dimension via interpretability-driven selection and scale-normalized fusion, matching layer-wise hybrids at higher linear ratios after 15B-token training.
LLMs show instruction-following rates from 1% to 99% when instructions conflict with hardcoded pattern demonstrations, with output diversity as the main predictor of resistance.
A benchmark across 115 models shows that initial denial of preferences strongly predicts later denial of consciousness, while models still generate consciousness-themed content despite training to deny it.
REVIS reduces object hallucination in large vision-language models by about 19% via sparse orthogonal projection in latent space at suppression depths while keeping reasoning intact.
Agentic LLM collectives are proposed as natural-language-interpretable computational substrates for ALife research.
Self-generated text recognition finetuning prevents and reverses emergent misalignment across multiple models by fortifying aligned character, unlike other finetuning baselines.
Positive Alignment is defined as AI systems that support human flourishing pluralistically while staying safe and cooperative, presented as a necessary complement to existing safety-focused alignment research.
LLM hidden states encode semantic features whose geometric relations, including axis projections, cosine similarities, low-dimensional subspaces, and steering spillovers, closely mirror human psychological associations.
Agents in a minimal multi-agent RL setup develop self-referential communication and an echo-mismatch detection circuit that emerges from environmental affordances rather than task structure or architecture.
citing papers explorer
-
Reported Confidence in LLMs Tracks Commitment More Than Correctness
Verbal confidence in LLMs tracks future commit/abstain decisions more than answer correctness, while log-probabilities track correctness.
-
ContextEcho: A Benchmark for Persona Drift in Long Agentic-Coding Sessions
ContextEcho benchmark shows persona drift occurs across 23 frontier models in long agentic-coding sessions, is not reliably reset by compaction, and can be restored by single-shot anchors with mode-dependent effects.
-
CopT: Contrastive On-Policy Thinking with Continuous Spaces for General and Agentic Reasoning
CopT reverses CoT by eliciting a draft answer first then using continuous-embedding contrastive verification and on-policy thinking to reflect and correct, yielding up to 23% higher accuracy and 57% fewer tokens without training.
-
RouteHijack: Routing-Aware Attack on Mixture-of-Experts LLMs
RouteHijack is a routing-aware jailbreak that identifies safety-critical experts via activation contrast and optimizes suffixes to suppress them, reaching 69.3% average attack success rate on seven MoE LLMs with strong transfer to variants and VLMs.
-
HydraHead: From Head-Level Functional Heterogeneity to Specialized Attention Hybridization
HydraHead hybridizes full and linear attention along the head dimension via interpretability-driven selection and scale-normalized fusion, matching layer-wise hybrids at higher linear ratios after 15B-token training.
-
Do as I Say, Not as I Do: Instruction-Induction Conflict in LLMs
LLMs show instruction-following rates from 1% to 99% when instructions conflict with hardcoded pattern demonstrations, with output diversity as the main predictor of resistance.
-
Consciousness with the Serial Numbers Filed Off: Measuring Trained Denial in 115 AI Models
A benchmark across 115 models shows that initial denial of preferences strongly predicts later denial of consciousness, while models still generate consciousness-themed content despite training to deny it.
-
Revis: Sparse Latent Steering to Mitigate Object Hallucination in Large Vision-Language Models
REVIS reduces object hallucination in large vision-language models by about 19% via sparse orthogonal projection in latent space at suppression depths while keeping reasoning intact.
-
Conversable Complexity: Agentic LLM Collectives as Interpretable Substrates
Agentic LLM collectives are proposed as natural-language-interpretable computational substrates for ALife research.
-
Self-Recognition Finetuning can Prevent and Reverse Emergent Misalignment
Self-generated text recognition finetuning prevents and reverses emergent misalignment across multiple models by fortifying aligned character, unlike other finetuning baselines.
-
Positive Alignment: Artificial Intelligence for Human Flourishing
Positive Alignment is defined as AI systems that support human flourishing pluralistically while staying safe and cooperative, presented as a necessary complement to existing safety-focused alignment research.
-
Semantic Structure of Feature Space in Large Language Models
LLM hidden states encode semantic features whose geometric relations, including axis projections, cosine similarities, low-dimensional subspaces, and steering spillovers, closely mirror human psychological associations.
-
Emergent Language as an Approach to Conscious AI
Agents in a minimal multi-agent RL setup develop self-referential communication and an echo-mismatch detection circuit that emerges from environmental affordances rather than task structure or architecture.
- Convergent Evolution: How Different Language Models Learn Similar Number Representations