Removed from main (all preserved on the dev branch): the arXiv build and
its sources, design documents (blueprint, results summary, review responses,
essay drafts), tasks/ and CLAUDE.md, the cover letter and reference tooling,
two unused manuscript figures, and every experiment that feeds no figure or
number in the paper: the collapse null, the sexual-vs-asexual lineage, the
NK speciation variant, the 0.5B single-seed LLM prototypes, the compose and
society experiments with their calibration and pilot runs, and their
configs, runners, tests, figure scripts and PBS jobs. Their result bundles
are moved to results/_archive/ (ignored) so the parquets stay on disk.
Also: plot_llm_speciation reads the s{seed}/ layout; the mating-breadth
plot writes under its bundle name; Makefile targets reduced to the kept
experiments; REPRODUCING.md and README point to dev for the rest.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Y64o8FKP7rCuXzC48pxpMm
Clarity pass over the main text (36-item audit), Discussion rewrite and cut,
acknowledgements, Souly et al. as ref 62, lettered SI panels, model section
moved under Results; plus the untracked curriculum/society/compose/smol
configs, runners, figures, stats and tests that the SI already cites.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Y64o8FKP7rCuXzC48pxpMm
Queue strategy per GG: one self-contained job per seed (founders inline - no
cache races, no dependencies), single L40S, 6 h walltime (measured: pilot ran
31 min/arm on the A4000, so a seed is ~2.5-3.5 h) - short independent jobs
that backfill well. Seed 1 runs locally on the A4000 as queue insurance;
seeds 2-4 go to CX3 as array elements. 0.5B for the campaign (sign-level,
seed-replicated, the paper's established pattern); a single 7B confirm of the
headline contrast is the post-campaign step.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BkRLcc18rwT2Lysu6PbG7v
Pilot v1: full and no_grounding indistinguishable - both arms pay a ~25%
distillation tax per generation and pin at the same mutation-selection floor,
so grounding has no differential to act on. v2 raises inheritance fidelity
(n_inherit 240->600, child epochs 2->3, founders 600x3) and adds two evenly-
applied knobs: elitism (top parent survives as an unmodified copy - overlapping
generations, compensating for lossy distillation where E11 had faithful
genotype copying) and n_parents (sharper truncation selection). v1 archived at
results/llm_society_pilot1; v2 running.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BkRLcc18rwT2Lysu6PbG7v
E11 re-instantiated in a population of LoRA agents, closing the paper's stated
gap before submission (GG: a weeks-scale experiment a reviewer would demand).
One grounding knob in the evaluation channel (g*verifier + (1-g)*conformity,
exactly E11); inheritance is identical in all arms and deliberately ungrounded
(children distilled from their source's own answers - self-consumption made
literal). Directed sex = complementary pairing + Dirichlet offspring screened
on the arm's own signal (the verifier never enters the no_grounding loop);
QD selection on verifier-free behavioural distance; terminal-degeneration
fallback copies the parent instead of crashing a sweep. Pure operators
unit-tested (155 green); smoke run end-to-end on the local A4000 already
shows the self-consumption signature (conformity up, diversity down in one
generation). Design, falsifiers, cost table: tasks/workorder-llm-society.md.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BkRLcc18rwT2Lysu6PbG7v
The decisive experiment from the external review. 39 LoRA parent pairs
(0.5B, 3 seeds) on three axes decorrelated by construction: conflict
(contradictory conventions on shared prompts, private budgets fixed),
compat (same prompts, SAME convention — overlap without conflict), and
duration (weight divergence, zero conflict). Six pre-merge predictors;
primary outcome = merge penalty (parent potential − merged achieved).
League table (Spearman vs penalty, n=39): functional measures predict
(dis_raw +0.460, epi_conf +0.446, p<0.005); geometry collapses
(delta_cos +0.03, delta_l2 +0.17 n.s.); gradient alignment weak (−0.35);
performance ~0. The first grid's apparent geometry win (+0.60) was an
overlap/volume artifact — the compat control axis (added for exactly
this) exposed and killed it: same overlap and data volume, zero penalty.
Honest riders in the README: confidence weighting does not beat raw
disagreement as a rank predictor (pre-registered internal prediction not
confirmed; it does double the conflict/compat level contrast), and
|rho|~0.45 is bounded by 0.5B merge-outcome noise (7B is the firm-up).
Also: micro-batched gradient accumulation (OOM fix on the shared 16GB
GPU), exact r-space LoRA-delta geometry (brute-force-verified test,
151 green), systemd-run runbook lesson (tmux dies with the SSH session
scope on this box).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BkRLcc18rwT2Lysu6PbG7v
llm_speciation (new kind; src/llm/speciation.py): E13 in LLM weights.
LoRA children share the frozen base's coordinates, so merge failure is
functional by construction. CONFLICT (ambiguous sort prompts learned
under opposite conventions — the BDM structure): function-specific
hybrid breakdown — merged coherence 0.02-0.08 falls below BOTH parents
(~0.2) on the conflicted function; and in the de-confounded `add` design
(private budget fixed, conflict added on top; 3 seeds after a
single-seed pilot showed one anomalous point) the merge's private-family
accuracy shows NO trend with conflict — the damage is surgical, not
global. DURATION (over-trained disjoint specialists, 1->12 epochs): the
merge improves (0.84->0.94) and stays above the best parent — the MLP
"no emergent isolation" null generalises; relevant to the
expert-training-duration report (2607.11997), with the epistasis
prediction left to the decisive experiment.
Multi-seed firm-up (seeds threaded into specialist caches; `seeds:` list
support in the runner; fixed test sets): all three recombination claims
hold with CIs — merges beat every specialist (5 seeds, ties
0.647±0.027 > best spec 0.592±0.009; worst-family 0.28 vs <=0.16); union
0.274±0.026 > fusion 0.174±0.102 on hard (3 seeds); directed 0.221±0.026
> soup. NEW finding: fusion is seed-FRAGILE where headroom exists
(CI ±0.10) while routing/directed selection are stable (±0.026) — the
union/selection operators win on reliability, not just mean.
Figures (llm_speciation 3-panel; llm_seeds 3-panel with 95% CI), READMEs,
+1 convention test (150 green), make llm-speciation / llm-seeds targets.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01BkRLcc18rwT2Lysu6PbG7v
The easy task families saturated 7B (strings & arith at 1.00), so the earlier
7B nulls — moe: fusion 0.87 > union 0.84; directed ~= soup — could not separate
"refinements don't help at scale" from "tasks too easy at 7B". Adds a hard task
variant (hard: true in tasks.py: multi-step lists, Caesar ciphers / letter
transforms, multi-step & larger arithmetic; same family labels and answer
formats, threaded through make_tasks/train_specialist/runners; hard specialists
cache separately as spec_*_hard) and re-runs both experiments at 7B on Imperial
CX3 (one L40S, 24 min, unsaturated: arith ~0.48, strings 0.67, lists 0.34).
Both nulls flip back to the 0.5B ordering:
- Union beats fusion again: routing 0.500 > fusion 0.40 (soup 0.392 / ties
0.400), the same 10-pt margin as 0.5B. Fusion dilutes the fragile strings
specialist so hard (0.665 -> soup 0.300) that soup even trails the best single
specialist (0.425); routing keeps it intact (0.670).
- Directed selection beats soup again: 0.492 > 0.392 (+10 pts), recovering most
of routing's benefit from one deployable merged model (lifts strings to 0.630).
Correction to the earlier interpretation: the llm_moe_hpc "regime flip" and the
llm_directed_hpc "no headroom" null were driven by TASK SATURATION, not base
capability. The operative variable is headroom — "merge, don't average" (union >
fusion) and "directed sex" (selection > single blend) hold whenever there is room
to lose to dilution: a weak base (0.5B) OR hard tasks at a strong base (7B-hard).
Fusion only wins in the degenerate corner where easy tasks let a strong base
compose to the 1.00 ceiling. Vindicates E8's max > mean in real 7B weights once
saturation is controlled.
Default (easy) task behaviour is unchanged (hard defaults False). +1 hard-task
test (131 green). Excludes the 0.5B smoke bundle (a pipeline gate, not a
deliverable). Results in results/llm_{moe,directed}_hard_hpc/ (parquet gitignored).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Adds the "directed sex" operator (E10) the moe regime-flip pointed to: don't
commit to one a-priori blend — breed a population of recombinant offspring
(specialists merged at Dirichlet-sampled weights), score each on a held-out
validation split with the verifier, and keep the fittest, reported on a fresh
test split. Two breeding objectives: best-overall and best-worst-family.
src/llm/directed.py + kind llm_directed, reusing the cached specialists.
Result — refinements pay off in proportion to how far the uniform soup is from
optimal:
- 0.5B (soup dilutes): directed selection beats soup on the bred objective —
directed_overall 0.69 > soup 0.64; directed_balanced worst-family 0.37 > 0.26.
Riders: single-objective selection trades off the other axis (overall-breed
tanks lists to 0.17); a global blend still trails per-input routing (0.74).
- 7B (Imperial CX3, soup already composes to ceiling on near-saturated families,
strings/arith 1.00): directed ~= soup (0.868 ~ 0.873, marginally below via a
val/test overfit gap) — no fitter offspring to breed.
Through-line across all four LLM runs: "merge, don't average" and its refinements
(routing, directed selection) are weak-base / suboptimal-default phenomena — they
help at 0.5B and are inert at 7B. Honest limitation kept in the writeup: the 7B
families are near-saturated, which caps the headroom; a harder unsaturated
benchmark is the fair next test.
Also folds in the two llm_moe local manifest/config files missed in 8da0dac.
+3 directed unit tests (130 green). Results in results/llm_directed{,_hpc}/
(parquet gitignored).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Adds the union-preserving recombination operator that llm_merge lacked (E8's max,
not mean): keep each specialist LoRA intact and SELECT the right one per prompt
(MoE router: oracle, or training-free nearest-centroid over base embeddings) or
per module (max_merge = winner-take-all by delta norm). src/llm/moe.py, kind
llm_moe, reuses the cached specialists.
Result — a clean regime boundary for "merge, don't average":
- 0.5B: union wins. Routing 0.74 / worst-family 0.43 > soup 0.64 / 0.26, with no
dilution (recovers each specialist's own-family peak). E8's max > mean in real
weights, because at a weak base averaging dilutes.
- 7B (Imperial CX3, L40S, 9 min): the ordering INVERTS. Fusion wins — soup 0.87 >
routing 0.84 > max_merge 0.78. Routing is capped at the best parent per family;
fusion blends and, given a capable base, COMPOSES beyond any parent (soup lists
0.62 > spec 0.57). Selection can't synthesise better than its best component;
averaging-that-composes can.
So "merge, don't average" (E4/E8) is a weak-parent / small-model law, not
universal: union wins under dilution, fusion wins under composition. Refines E8
(its additive-landscape max>mean assumed no compositional headroom). The operator
to want is fusion-that-composes + offspring selection = the directed-sex ideal
(E10) — the natural next experiment.
Honest riders: the learned router is trivially perfect (lexically-distinct
families), and router-free max_merge is the weakest union (not input-adaptive).
+2 router unit tests (127 green). Results in results/llm_moe{,_hpc}/ (parquet
gitignored per the reproducibility contract).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Enable running the LLM tier on Imperial's HPC (scheduler: PBS Pro / qsub).
- hpc/probe.pbs: 10-min 1-GPU reconnaissance job resolving the two unknowns
the RCS docs omit -- compute-node internet access and the L40S driver's
CUDA version -- plus TMPDIR/disk and available python/cuda modules.
- hpc/llm_merge.pbs: scaled run on an L40S (48 GB), offline HF-cache wired,
runs configs/llm/merge_hpc.yaml.
- configs/llm/merge_hpc.yaml: Qwen2.5-7B-Instruct (fits the L40S) to reduce
the noise that left the 0.5B prototype's overall-exceeds sign marginal.
- hpc/README.md: the git-based workflow (login-node uv env + model
pre-download -> qsub -> rsync results back), confirmed PBS/GPU directives,
and the code TODOs for the definitive run (multi-seed, more families,
directed/dilution-resistant merge).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
First step from toy models toward real language models, on one 16 GB GPU.
New src/llm/ package: procedural task families + exact-match verifier
(tasks.py), batched eval (evaluate.py), LoRA specialisation (specialise.py,
manual answer-only SFT), weight-space merge via peft add_weighted_adapter
(merge.py: soup = averaged deltas, ties = sign-reconciled union), runner
(experiment.py, kind llm_merge). Base Qwen2.5-0.5B-Instruct (Apache-2.0);
three disjoint hard families (lists/strings/arith); one LoRA specialist each
(~90s total).
Result (seed 1), reported honestly:
- STRONG/robust: the merges are the ONLY models competent across ALL
families -- worst-family ~0.25 vs <0.16 for every single specialist (the
Fisher-Muller "generalist assembled from specialists" signature, in real
LoRA weights).
- MARGINAL: "exceeds every parent overall" is only marginal at this scale
(soup 0.64 vs best specialist 0.63; ties 0.61 below it).
- CAVEAT VISIBLE: averaging dilutes peaks (lists specialist 0.43 -> merge
0.26) -- Layer-1's "merge, don't average" (E4) appearing in real weights.
The pipeline works end-to-end; the balance/retention half reproduces; the
strict overall-exceeds and soup-vs-ties distinction need scale (bigger base,
more/cleaner families, seeds, a dilution-resistant / offspring-selected
merge) -- the HPC step. Env: Python 3.14 + transformers 5.13 works;
note transformers-5.x apply_chat_template returns a dict. make env-llm /
make llm; adapters under gitignored models/llm/, base in the HF cache.
figures/plot_llm_merge.py, README, tests/test_llm.py (+3, 125 green).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
