# GT information regimes, KET, and TopoCoend

- **Registry ID:** `catagi-information-ket`
- **Book:** Categories for AGI
- **Documentation status:** `curation-required`
- **Experimental status:** controlled comparisons with explicit leakage controls
- **Canonical documentation URL:** https://categorical-ai.sridharmahadevan.com/experiments/catagi-information-ket

## Purpose

Strict-causal and future-carrier language-modeling comparisons on PTB and WikiText corpora.

## Book location

- Chapter 9: Information Regimes in Geometric Transformers
- Chapter 10: Kan Extension and Topological Coend Transformers
- Chapter 11: Structured Language Modeling

## Documented studies

- Strict-causal versus future-carrier leakage control
- PTB blockwise KET comparison
- WikiText-2 KET diagnostics
- WikiText-103 KET and TopoCoend comparison

## Associated code packages

- `categories-gt-archive` — Geometric and Kan Extension Transformer archive; **local-not-public**; license decision pending. Working scripts and result directories for DB, GT, KET, language modeling, manifold, mean-field, and scaling studies. A curated public packet remains to be extracted.
- `lincs-allora` — [LINCS ALLORA and LINCS-KET package](https://github.com/sridharmahadevan/LINCS/tree/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora); **public**; Apache-2.0. Adapter-composition and KET repair implementations with smoke and extended experiment entry points.

## Start here

These concrete files are selected from the complete resolved source surface. They orient the reader; they are not a claim that every family is independently reproducible.

- **runner · staged-not-public** — `categories-gt-archive:KET/ket_experiments/eval/eval_ppl.py`. Locate this file in the curated companion packet; no public download is currently offered.
- **implementation · public-pinned** — [`lincs-allora:KET/aggregate_gpt2_cuda_replication.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/aggregate_gpt2_cuda_replication.py). Open the pinned public source file.
- **runner · staged-not-public** — `categories-gt-archive:KET/ket_experiments/eval/freeze_ablate.py`. Locate this file in the curated companion packet; no public download is currently offered.
- **runner · staged-not-public** — `categories-gt-archive:KET/ket_experiments/eval/model_utility.py`. Locate this file in the curated companion packet; no public download is currently offered.

## Entry points

- `KET/ket_experiments/`
- `scripts/ket_experiments/`

## Result artifacts

- `loss/perplexity plots and recorded comparison tables`

## Resolved code surfaces (42)

- **dependency** — `categories-gt-archive:KET/ket_experiments/datasets_ptb_wt.py`
- **runner** — `categories-gt-archive:KET/ket_experiments/eval/__init__.py`
- **runner** — `categories-gt-archive:KET/ket_experiments/eval/eval_ppl.py`
- **runner** — `categories-gt-archive:KET/ket_experiments/eval/freeze_ablate.py`
- **runner** — `categories-gt-archive:KET/ket_experiments/eval/model_utility.py`
- **runner** — `categories-gt-archive:KET/ket_experiments/eval/ngram_sweep.py`
- **runner** — `categories-gt-archive:KET/ket_experiments/eval/perm_ablate.py`
- **runner** — `categories-gt-archive:KET/ket_experiments/ket_unified_harness.py`
- **implementation** — `categories-gt-archive:KET/ket_experiments/models/__init__.py`
- **implementation** — `categories-gt-archive:KET/ket_experiments/models/block_heads.py`
- **implementation** — `categories-gt-archive:KET/ket_experiments/models/ket_block_wrappers.py`
- **implementation** — `categories-gt-archive:KET/ket_experiments/models/ket_model.py`
- **runner** — `categories-gt-archive:KET/ket_experiments/train/train_ket.py`
- **runner** — `categories-gt-archive:KET/ket_experiments/train/train_ket_block.py`
- **dependency** — `categories-gt-archive:KET/ket_experiments/train_torch_gt_lm.py`
- **dependency** — `categories-gt-archive:KET/ket_experiments/train_torch_kan_lm_ptbwt_incidence_mp_compat.py`
- **dependency** — `categories-gt-archive:KET/ket_experiments/train_torch_kan_lm_ptbwt_softpred_hint.py`
- **dependency** — `categories-gt-archive:KET/ket_experiments/train_torch_lm.py`
- **dependency** — `categories-gt-archive:KET/ket_experiments/train_torch_topocoend_lm_ptbwt_compat.py`
- **implementation** — `categories-gt-archive:KET/ket_experiments/training/block_utils.py`
- **implementation** — [`lincs-allora:KET/__init__.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/__init__.py)
- **implementation** — [`lincs-allora:KET/aggregate_gpt2_cuda_replication.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/aggregate_gpt2_cuda_replication.py)
- **implementation** — [`lincs-allora:KET/aggregate_model_zoo.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/aggregate_model_zoo.py)
- **implementation** — [`lincs-allora:KET/allora_ket_adapters.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/allora_ket_adapters.py)
- **implementation** — [`lincs-allora:KET/analyze_adapter_sweep.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/analyze_adapter_sweep.py)
- **implementation** — [`lincs-allora:KET/analyze_capability_safety.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/analyze_capability_safety.py)
- **implementation** — [`lincs-allora:KET/analyze_confirmatory.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/analyze_confirmatory.py)
- **implementation** — [`lincs-allora:KET/analyze_lora_text_sweep.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/analyze_lora_text_sweep.py)
- **implementation** — [`lincs-allora:KET/analyze_theory_diagnostics.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/analyze_theory_diagnostics.py)
- **implementation** — [`lincs-allora:KET/capability_safety_composition.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/capability_safety_composition.py)
- **implementation** — [`lincs-allora:KET/cross_fitted_quotient_prediction.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/cross_fitted_quotient_prediction.py)
- **implementation** — [`lincs-allora:KET/factor_quotient_projectability.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/factor_quotient_projectability.py)
- **implementation** — [`lincs-allora:KET/gpt2_medium_expanded_eval.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/gpt2_medium_expanded_eval.py)
- **implementation** — [`lincs-allora:KET/gpt2_medium_mps_pilot.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/gpt2_medium_mps_pilot.py)
- **implementation** — [`lincs-allora:KET/gpt2_sequential_lora_smoke.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/gpt2_sequential_lora_smoke.py)
- **implementation** — [`lincs-allora:KET/ket_adapter_smoke.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/ket_adapter_smoke.py)
- **implementation** — [`lincs-allora:KET/ket_coend_comparison.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/ket_coend_comparison.py)
- **implementation** — [`lincs-allora:KET/ket_ptb_adapter_train.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/ket_ptb_adapter_train.py)
- **implementation** — [`lincs-allora:KET/lora_text_source_smoke.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/lora_text_source_smoke.py)
- **implementation** — [`lincs-allora:KET/model_registry.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/model_registry.py)
- **implementation** — [`lincs-allora:KET/reflective_allora_order.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/reflective_allora_order.py)
- **implementation** — [`lincs-allora:KET/stochastic_quotient_aggregation.py`](https://github.com/sridharmahadevan/LINCS/blob/c78bfe50ded8d40db90c614e40dd2e0e146a9875/packages/allora/KET/stochastic_quotient_aggregation.py)

## Frozen run records (0)

- No frozen run-level record is registered. Consult the result-artifact list and evidence boundary above for the exact surviving evidence; absence of a run record is not evidence that no experiment was run.

## Evidence boundary

Future-carrier runs are leakage controls, not valid autoregressive improvements; point estimates are not population effects.

A code or artifact link establishes traceability. It does not by itself establish independent reproduction, statistical adequacy, correctness, or support for a claim beyond this boundary.