mlsanityData to 14 Oct 2026

The Ignition Index: Measuring Global Workspace Dynamics in Language Models

arXiv 2608.05160 ↗First posted 26 May 2026cs.AI cs.CL cs.LGcited by 0

On the map: Mechanistic interpretability → Circuits and mechanistic analysis

Abstract (arXiv)

We introduce the Ignition Index (I), a validated scalar metric that operationalizes Global Workspace Theory's (GWT) all-or-none ignition prediction in transformer language models. The metric fits a four-parameter sigmoid to per-layer linear probe accuracy as a function of input signal strength, extracting steepness parameter beta-hat: high values indicate abrupt, ignition-like transitions; low values indicate graded build-up. Across 11 models spanning five architecture families, shuffled-label controls demonstrate 9.6-fold selectivity for genuine linguistic structure over spurious probe capacity (p < 0.001, Mann-Whitney U-test). We find: (1) Feedforward transformers exceed SSMs by 89% in aggregate beta-hat (p < 1e-13, Cohen's d = 0.52), with Mamba exhibiting near-linear profiles consistent with absent global broadcast. (2) Huginn-3.5B exhibits 2.12-fold higher ignition along its iteration axis than its depth axis, demonstrating that recurrent architectures manifest workspace-like transitions along the recurrence dimension. (3) Pythia-410M shows a PELT-detected phase transition at training step 256 (+67%), preceding induction-head formation. (4) Hypotheses linking ignition to model scale and signal strength were not confirmed, suggesting transformer architectures may saturate available ignition mechanisms. The Ignition Index provides the first validated quantitative bridge between GWT's dynamical predictions and mechanistic interpretability, with 9.6-fold measurement selectivity and architecture-level discriminability not previously characterized in the scaling literature. Code: https://github.com/saman-rahbar/ignition-index

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Circuits and mechanistic analysis, one of 13 directions in Mechanistic interpretability. Builds on 25 of its 62 references that are in the map. First posted on arXiv 26 May 2026.

Builds on

25 of 62 references.

Closest ancestors · 3

Same direction, newest first

  1. 2023Towards Automated Circuit Discovery for Mechanistic Interpretability28 Apr
  2. 2023Progress measures for grokking via mechanistic interpretability12 Jan
  3. 2022Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small1 Nov

Other references · 22

  1. 2023Mamba: Linear-Time Sequence Modeling with Selective State Spacescited by 4,718
  2. 2022Locating and Editing Factual Associations in GPTcited by 2,691
  3. 2023Sparse Autoencoders Find Highly Interpretable Features in Language Modelscited by 1,420
  4. 2016Understanding intermediate layers using linear classifier probescited by 1,353
  5. 2019BERT Rediscovers the Classical NLP Pipelinecited by 1,059
  6. 2022Toy Models of Superpositioncited by 971
  7. 2021Probing Classifiers: Promises, Shortcomings, and Advancescited by 876
  8. 2018Universal Transformerscited by 647
  9. 2019Designing and Interpreting Probes with Control Taskscited by 640
  10. 2023Eliciting Latent Predictions from Transformers with the Tuned Lenscited by 623
  11. 2019BLiMP: The Benchmark of Linguistic Minimal Pairs for Englishcited by 564
  12. 2023Are Emergent Abilities of Large Language Models a Mirage?cited by 523
10 more
  1. 2024Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Modelscited by 435
  2. 2025Scaling up Test-Time Compute with Latent Reasoning: A Recurrent Depth Approachcited by 415
  3. 2024Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2cited by 394
  4. 2020Inductive Biases for Deep Learning of Higher-Level Cognitioncited by 275
  5. 2020COGS: A Compositional Generalization Challenge Based on Semantic Interpretationcited by 263
  6. 2020Information-Theoretic Probing with Minimum Description Lengthcited by 232
  7. 2023Finding Alignments Between Interpretable Causal Variables and Distributed Neural Representationscited by 209
  8. 2017The Consciousness Priorcited by 107
  9. 2023Consciousness in Artificial Intelligence: Insights from the Science of Consciousnesscited by 104
  10. 2020Deep Learning and the Global Workspace Theorycited by 36

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026Residual Streams Read, Recurrent States Remember: The Global Workspace in Mamba Models
  2. 2024The Expressive Capacity of State Space Models: A Formal Language Perspective
  3. 2025Scaling Laws for State Dynamics in Large Language Models
  4. 2022minicons: Enabling Flexible Behavioral and Representational Analyses of Transformer Language Models
  5. 2026Interpreting and Steering State-Space Models via Activation Subspace Bottlenecks
  6. 2026FishBack: Pullback Fisher Geometry for Optimal Activation Steering in Transformers
  7. 2026Hallucination as Trajectory Commitment: Causal Evidence for Asymmetric Attractor Dynamics in Transformer Generation
  8. 2025Three tiers of computation in transformers and in brain architectures