Skip to content
No results
  • Explainers
J-Space logo
J-Space
  • Explainers
J-Space logo
J-Space
  • Deep Dives

Knowing Both Facts Still Fails the Composed Two-Hop Query

Two-hop circuits compose facts via a residual-stream bridge instead of an A-to-C shortcut. Knowing both facts still fails the composed query.

  • Editorial Team
  • September 11, 2026
  • Deep Dives

Harmful Demo Count Flips Refusal Before SAE Features Disappear

Packing many harmful demos into long context lifts attack success. Refusal features may stay SAE-readable even when the model complies.

  • Editorial Team
  • September 10, 2026
  • AI Safety

Residual Jacobians Reveal Misalignment That Sampled Tokens Still Hide

Sampled tokens can look aligned while residual Jacobians have already reoriented. Run JVP estimators and a five-stage protocol for NIST and EU GPAI logs.

  • Editorial Team
  • September 9, 2026
  • Philosophy of AI

Call It a Workspace Only After Nonlinear Ignition and Capacity Limits

Global workspace theory ignites then broadcasts under capacity limits. Scaled dot-product attention only routes over token positions.

  • Editorial Team
  • September 8, 2026
  • Deep Dives

The Shared Bus Lives in Residual Streams, Not Attention Maps

Residual streams analogize capacity-limited broadcast better than attention maps. Softmax routing stays graded and content-addressable, not ignited.

  • Editorial Team
  • September 3, 2026
  • Deep Dives

Softmax Weights Replace the Ignition Threshold Global Workspace Theory Needs

GWT treats conscious access as competition for a limited broadcast. Transformers use graded QKV routing on a residual stream, not ignition.

  • Editorial Team
  • September 2, 2026
  • Philosophy of AI

Decoder-Only Transformers Route Tokens Instead of Igniting a Global Workspace

QKV attention routes tokens rather than igniting broadcast. Decoder-only transformers fail GWT; the 2017 paper dropped recurrence. Use indicator tests.

  • Editorial Team
  • September 1, 2026
  • Deep Dives

Claude Residual Streams Need 34 Million SAE Features, Not Chat Logs

Decode residual-stream activations and 34 million SAE features; chat transcripts, chain of thought, and the public API do not expose them.

  • Editorial Team
  • August 31, 2026
  • Philosophy of AI

Neither Recurrent Loops Nor Untied Depth Give Transformers Self-Monitoring

GPT-3 stacks 175 billion untied parameters while Universal Transformers reuse one block. Both broadcast. Neither creates C2 sentience.

  • Editorial Team
  • August 26, 2026
  • Deep Dives

Unembedding Proximity Pulls Residual Effects Into Late Layers

Late residual writes sit immediately upstream of the unembedding and bias decoder-aligned effects toward the last blocks.

  • Editorial Team
  • August 25, 2026
1 2 3
Next
Copyright © 2026 - WordPress Theme by CreativeThemes