FLIPS identifies LLM instances with 96% closed-set and 90% open-set accuracy by exploiting biases in generated binary random sequences across 237 instances.
Title resolution pending
9 Pith papers cite this work. Polarity classification is still indexing.
9
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
years
2026 9verdicts
UNVERDICTED 9roles
background 1polarities
background 1representative citing papers
A Riemannian geodesic framework for label-free manifold steering in language models via a schema-supervised encoder approximating output Hellinger distance on activations.
Linear probes for Othello board states factor into tensor-product structure with square and color embeddings composed by a binding matrix, from which the linear probes can be directly recovered.
State-writing models causally use edited scratchpad states in a controlled task at 80-91% accuracy on held-out examples, unlike final-answer-only and pretrained controls.
Co-activation clustering of attention heads proposes candidate circuits that pass causal closure validation in dense 1B models but fail in a Mixture-of-Experts model, where ablation can improve loss.
Linear probes recover day-of-year from LM activations for temporal reasoning but are orthogonal to the model's causal 4D subspace identified by DAS, with the angle matching the Haar-uniform random null, replicated across scales and families.
Diverse language models converge on similar periodic number features with a two-tier hierarchy of Fourier sparsity and geometric separability, acquired via language co-occurrences or multi-token arithmetic.
Causal localization via attribution and patching identifies a temporal preference subgraph in mid-to-upper layers of Qwen3-4B-Instruct-2507, with time-horizon geometry in the residual stream and initial evidence for steering-vector control.
H-probes locate low-dimensional subspaces encoding hierarchy in LLM activations for synthetic tree tasks, show causal importance and generalization, and detect weaker signals in mathematical reasoning traces.
citing papers explorer
-
FLIPS: Instance-Fingerprinting for LLMs via Pseudo-random Sequences
FLIPS identifies LLM instances with 96% closed-set and 90% open-set accuracy by exploiting biases in generated binary random sequences across 237 instances.
-
Riemannian-Manifold Steering: Geometry-Aware Generative Autoencoders for Label-Free Steering
A Riemannian geodesic framework for label-free manifold steering in language models via a schema-supervised encoder approximating output Hellinger distance on activations.
-
Tensor Product Representation Probes Reveal Shared Structure Across Linear Directions
Linear probes for Othello board states factor into tensor-product structure with square and color embeddings composed by a binding matrix, from which the linear probes can be directly recovered.
-
Do Models Read What They Write? Causal Registers in Scratchpad Reasoning
State-writing models causally use edited scratchpad states in a controlled task at 80-91% accuracy on held-out examples, unlike final-answer-only and pretrained controls.
-
Closure-Validated Circuit Discovery in Attention Heads: Co-activation Proposes, Ablation Disposes
Co-activation clustering of attention heads proposes candidate circuits that pass causal closure validation in dense 1B models but fail in a Mixture-of-Experts model, where ablation can improve loss.
-
When and How Long? The Readout-Mediator Angle in Temporal Reasoning
Linear probes recover day-of-year from LM activations for temporal reasoning but are orthogonal to the model's causal 4D subspace identified by DAS, with the angle matching the Haar-uniform random null, replicated across scales and families.
-
Temporal Preference Concepts and their Functions in a Large Language Model
Causal localization via attribution and patching identifies a temporal preference subgraph in mid-to-upper layers of Qwen3-4B-Instruct-2507, with time-horizon geometry in the residual stream and initial evidence for steering-vector control.