Sep 9

1. 03:12

   CLERK paper draft v0.1: method, benchmark spec, protocol; results tables pending observed runs

   Falln87
   Verified signature

   76a8e7d
2. 03:10

   CLERK project README: method summary, novelty statement, repro instructions

   Falln87
   Verified signature

   7aab365
3. 03:10

   CLERK end-to-end reproducible pipeline (CPU smoke tests -> data -> LoRA SFT -> both evals)

   Falln87
   Verified signature

   1c428b7
4. 03:09

   eval_locomo: fix choice_logprobs continuation scoring

   Falln87
   Verified signature

   920a0c6
5. 03:09

   eval_locomo: clean rewrite (working choice log-likelihood scorer, consistent templates, single model loader)

   Falln87
   Verified signature

   705d9d4
6. 03:08

   CLERK LoCoMo-MC10 zero-shot eval: incremental ledger build over LoCoMo sessions, MC log-likelihood scoring, full-context + prompted-ledger + summary-RAG baselines

   Falln87
   Verified signature

   ca30e35
7. 03:07

   generator: fix gold_ops_for_events double-application bug; single-state implementation

   Falln87
   Verified signature

   1d30dff
8. 03:05

   tests_smoke: clean rewrite of invariant checks (gold ops executable, budget, round-trip, QA consistency)

   Falln87
   Verified signature

   76e7c09
9. 03:04

   eval_synthetic: initialize stale-error counters in Runner.__init__

   Falln87
   Verified signature

   c88c20f
10. 03:03

    CLERK pinned dependencies

    Falln87
    Verified signature

    79f4cda
11. 03:03

    CLERK smoke tests: gold ops executable, ledger round-trip, budget enforcement, QA supervision consistency

    Falln87
    Verified signature

    415817c
12. 03:02

    eval_synthetic: full clean rewrite (consistent generate contract, main, serialization helper)

    Falln87
    Verified signature

    f8cde5f
13. 03:02

    CLERK synthetic eval: clerk rollout + prompted-ledger/full-context/sliding-window/RAG baselines, per-type accuracy, memory-state metrics, token costs

    Falln87
    Verified signature

    193e1ae
14. 03:00

    train_sft: honor --base-model and --max-length, drop unused import

    Falln87
    Verified signature

    6380390
15. 03:00

    CLERK LoRA SFT trainer (TRL SFTTrainer), verified against trl examples/sft_qlora; pushes adapter to the Hub

    Falln87
    Verified signature

    4d8daa4
16. 03:00

    CLERK SFT dataset builder: timelines -> messages JSONL (CONSOLIDATE + ANSWER mixture), seeded shuffle

    Falln87
    Verified signature

    c498706
17. 03:00

    CLERK shared prompt definitions (train/eval must never drift)

    Falln87
    Verified signature

    865342a
18. 02:59

    generator: temporal answer uses slot-held value; gold ADD evicts first when ledger full; budget default 12 for eviction pressure

    Falln87
    Verified signature

    9dae931
19. 02:58

    generator: remove module-global session index; question targets restricted to ledger-supported predicates

    Falln87
    Verified signature

    61cccdd
20. 02:57

    common.py: fix facts_of valid_only semantics

    Falln87
    Verified signature

    a042d28
21. 02:57

    CLERK generator: correct single-pass timeline builder, ledger-aware question targets, eviction-safe gold ops

    Falln87
    Verified signature

    d03cb0b
22. 02:56

    CLERK generator: clean full rewrite of render_session and persona timeline builder

    Falln87
    Verified signature

    3cd7b4e
23. 02:55

    CLERK synthetic evolving-session generator: personas, updates/supersessions/tombstones, gold ledger transitions, seeded

    Falln87
    Verified signature

    fdcd0a4
24. 02:55

    CLERK core: slot schema, edit-op schema, deterministic reducer, serializers

    Falln87
    Verified signature

    4f70a26
25. 02:54

    initial commit

    Falln87
    Verified signature

    36958cf