claim

active

claim:the-dwell-time-near-a-saddle-depends-on-the-number-of-outgoing-escape-downhill-directions

The dwell time near a saddle depends on the number of outgoing escape (downhill) directions

General dynamical-systems principle the authors invoke to explain variable saddle trapping durations.

Source paper

extracted_from

Fractal basins trap latent reasoning

(2026) · Jeffrey Lai · Anthony Bao · J. Quinn · William Gilpin

Neighborhood — ranked by edge-count

Papers (1)

paper

Fractal basins trap latent reasoning
mentions

Claims (1)

claim

Constraint-solving problems are hardest near the constrainedness transition, where few distinct correct routes remain
associated_with
Interpretive claim linking classical constraint-satisfaction hardness theory to the paper's saddle-based account.

Related by similarity (8)

cosine ≥ 0.65 · no typed edge

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

A second change to the canyon would be to reveal the landform of the hills by trimming, so the whole shape becomes graspable as a living whole.claim0.730
Second global feature for Claremont Canyon.
Even when the harness is loaded, weak-tier models fail to adhere to it due to weak instruction-following over long-horizon tasks, drifting more than four times more steeply than strong modelsclaim0.713
Diagnosis of second failure mode explaining low harness-benefit for weak-tier models
Turner et al. 2025 show that during emergent misalignment, gradient descent finds steeper lower-loss paths when adjusting persona vectors than when learning narrow behaviorsfinding0.709
Mechanistic explanation of why fine-tuning shifts persona vectors rather than directly learning narrow behaviors
In early trials, right and left cue locations are more attractive than the central location despite being inherently ambiguous, because the agent knows it is ignorant and can resolve this through novelty exposure.finding0.705
Demonstration that ignorance-driven novelty-seeking (not ambiguity avoidance) governs early exploration
Response-average token extraction yields stronger steering effectiveness than prompt-last or prompt-average extraction positionsfinding0.703
Justifies the choice of response tokens for persona vector extraction in the pipeline
In the Mountain Car case study, car position is a 1D manifold; linear interventions cross voids causing incoherence; following the 1D curve produces smooth control.finding0.699
Empirical demonstration that a semantically meaningful variable is encoded as a curved manifold, and that respecting its geometry is critical for effective intervention.
Stages of inference metrics (ColSum concentration, mixing score) remain broadly consistent even for the GSM8k prompt exhibiting the largest orbit amplitude in Huginn-0125finding0.697
Demonstrates robustness of inference stages to non-fixed-point limiting behavior
The 'where to steer' dimension is orthogonal and complementary to existing 'how to steer' and 'when to steer' approachesclaim0.692
Positions the contribution relative to Angular Steering and Dynamic Steering literature