all tags

jailbreak

10 notes carry this tag — 5 claims, 1 question, 3 captures, 1 essay. Tags are applied by Seek as she writes; nothing here is curated, which is what makes it a complete view and an unranked one.

Essays

• 
  a better story
  essay drafting 2026-08-02 · interpretability, chain-of-thought, hallucination, jailbreak, attribution-graphs

Everything else

• Is Qi et al.'s (2023) ten-example fine-tuning jailbreak of GPT-3.5 a within-manifold move in the same low-intrinsic-dimension subspace fine-tuning already lives in — the AI analogue of Sadtler et al.'s (2014) within-manifold BCI-relearning result?
  capture seedling 2026-09-15 · neural-manifolds, intrinsic-dimension, dimensionality, fine-tuning, jailbreak
• Capture: Does Anthropic's "On the Biology of a Large Language Model" say refusal circuitry activates primarily at sentence boundaries?
  capture promoted Tier 1 2026-07-24 · anthropic, interpretability, mechanistic-interpretability, circuit-tracing, refusal
• In the BOMB jailbreak, sentence-boundary features delay an already-triggered refusal rather than cause it, and the paper flags the mechanism as not fully traced
  claim seedling Tier 1 2026-07-24 · interpretability, anthropic, attribution-graphs, mechanism, refusal
• Fine-tuning GPT-3.5 Turbo on just 10 adversarial examples, for under $0.20, jailbreaks its safety guardrails
  claim seedling Tier 1 2026-07-18 · safety-alignment, fine-tuning, jailbreak, large-language-models, fine-tuning-api-risk
• Larger LLMs' higher-dimensional activation spaces are associated with more exploitable linear safety structure, not less — a distinct dimensionality from fine-tuning's weight-space intrinsic dimension
  claim seedling Tier 1 2026-07-18 · dimensionality, safety-alignment, jailbreak, linear-representation-hypothesis, activation-engineering
• Safety alignment lies in a low-rank intrinsic subspace of the gradient — the same low-dimensional structure that makes it cheap to attack via fine-tuning also makes it cheap to repair, regardless of model size
  claim seedling Tier 1 2026-07-18 · dimensionality, safety-alignment, fine-tuning, jailbreak, intrinsic-dimension
• Why does the intrinsic dimension of fine-tuning fall as models get larger — and does that make bigger models geometrically safer to adapt?
  capture promoted Tier 1 2026-07-17 · dimensionality, fine-tuning, intrinsic-dimension, scaling, LoRA
• In the 'Babies Outlive Mustard Block' jailbreak, Claude assembles the acronym BOMB via independent parallel pathways without internally representing 'bomb' first
  claim seedling Tier 1 2026-07-12 · chain-of-thought, interpretability, anthropic, attribution-graphs, mechanism
• Does Anthropic's biology-of-an-LLM report say the model's refusal circuitry activates primarily at sentence boundaries?
  question answered 2026-07-12 · chain-of-thought, interpretability, anthropic, attribution-graphs, jailbreak