record state
frontier-ownedReview status
This finding is part of accepted frontier state. Review events, reviewable changes, and proof state explain how it can change.
frontiers / frontier
Finding bundle
back to stateno incoming links yet
record state
frontier-ownedThis finding is part of accepted frontier state. Review events, reviewable changes, and proof state explain how it can change.
finding statement
finding typeNo entity list is declared.
evidence
source-boundcomputational · manual state transition
proof impact
packet context2 reviewable changes and 0 evaluation records are attached to this finding id.
Evidence and conditions
method
manual state transition
evidence type
computational
conditions
Provenance
source title
mechinterp causal sweep
authors
agent:replicator
Across all five transformers (gpt2, distilgpt2, pythia-70m, pythia-160m, gpt2-medium), the in-context-repetition circuit is universal and stereotyped by depth: duplicate-token heads emerge shallowest (layers 0-1), previous-token heads at early-to-mid depth, and induction heads in the middle-to-deep band (~30-90% relative depth); the induction-head group is causally necessary in every model (group-ablation delta 4.95-11.52 nats, all per-seed deltas positive).
vs_2a6614cf8c092119 · manual_curation
outgoing
No outgoing links.
incoming
No incoming links.
events
vev_0e2ce4e5ffa52f6cfinding.assertedManual finding added to frontier state
agent:replicator · 2026-05-29
vev_833001ee0ddf660cfinding.caveatedVerifier caution: The core claim—that a group of induction heads is causally necessary across 5 models—is valid and well-supported by the repeated-token ablation. However, the language significantly overstates the scope: 'universal' and 'stereotyped by depth' imply properties across tasks, archite
reviewer:will-blair · 2026-05-29
reviewable changes
vpr_23db2dbfd9b1e369finding.addManual finding added to frontier state
applied · agent:replicator · 2026-05-29
vpr_3906611ccc2f44e6finding.caveatVerifier caution: The core claim—that a group of induction heads is causally necessary across 5 models—is valid and well-supported by the repeated-token ablation. However, the language significantly overstates the scope: 'universal' and 'stereotyped by depth' imply properties across tasks, archite
applied · reviewer:will-blair · 2026-05-29
evaluations
No evaluation record targets this finding id.