In a stochastic k-ary tree, a two-head transformer learns randomized DFS via policy gradient under depth-wise curriculum, generalizes to deeper trees, and adapts to imbalanced goals via discounting.
How to think step-by-step: A mechanistic understanding of chain-of-thought reason- ing
7 Pith papers cite this work. Polarity classification is still indexing.
7
Pith papers citing it
representative citing papers
In 1-3B instruction-tuned LMs on GSM8K, arithmetic CoT readout is dominated by positional copying of the trailing number before the answer delimiter, accounting for 54-92 percentage points of accuracy.
Counterfactual prompting effects on LLMs are often indistinguishable from those caused by meaning-preserving paraphrases, causing most previously reported demographic sensitivities to disappear under proper statistical comparison.
RLVR drives a concept network in LLMs through nucleation and freezing into inverse trees that support slow thinking, and intervening with brief SFT at peak frustration outperforms standard RLVR while post-freeze SFT causes forgetting.
Sparse autoencoders plus greedy filtering and factorization-machine interaction modeling identify minimal sets of features in Gemma-2-2B-IT and LLaMA-3.1-8B-IT whose ablation produces jailbreaks by flipping refusal to compliance.
ShadowCoT introduces a reasoning-level backdoor attack on LLMs achieving 94.4% attack success rate and 88.4% hijacking success rate with 0.15% parameter updates via internal state conditioning and reasoning chain pollution.
Reasoning in large output spaces proceeds via shortlisting then fine-grained reasoning; this characterization enables a mechanistic distillation strategy that outperforms standard distillation.
citing papers explorer
-
Emergent Slow Thinking in LLMs as Inverse Tree Freezing
RLVR drives a concept network in LLMs through nucleation and freezing into inverse trees that support slow thinking, and intervening with brief SFT at peak frustration outperforms standard RLVR while post-freeze SFT causes forgetting.
-
Beyond I'm Sorry, I Can't: Dissecting Large Language Model Refusal
Sparse autoencoders plus greedy filtering and factorization-machine interaction modeling identify minimal sets of features in Gemma-2-2B-IT and LLaMA-3.1-8B-IT whose ablation produces jailbreaks by flipping refusal to compliance.
-
ShadowCoT: Cognitive Hijacking for Stealthy Reasoning Backdoors in LLMs
ShadowCoT introduces a reasoning-level backdoor attack on LLMs achieving 94.4% attack success rate and 88.4% hijacking success rate with 0.15% parameter updates via internal state conditioning and reasoning chain pollution.