File size: 1,924 Bytes
53525ef
455fe3a
53525ef
455fe3a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
53525ef
455fe3a
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
---
pretty_name: KLIK Temporal Memory Paper
license: cc-by-4.0
task_categories:
  - other
language:
  - en
tags:
  - research-paper
  - agent-memory
  - temporal-memory
  - entity-aware-retrieval
  - privacy
  - evaluation-protocol
size_categories:
  - n<1K
configs:
  - config_name: publication-record
    data_files:
      - split: train
        path: paper.jsonl
---

# KLIK Temporal Memory Paper

This dataset is the public research record for **KLIK Temporal, Entity-Aware, Privacy-Constrained Memory**. It packages the public-edition manuscript, a citation record, and a machine-readable publication entry.

## What this release is

- A scoped architecture proposal for temporal, entity-aware, privacy-constrained agent memory.
- A preregistered protocol for comparing the proposed system with registered controls and memory baselines.
- An evidence-readiness audit that keeps historical retrieval observations separate from unmeasured target-architecture results.

## What this release is not

- Not a claim that the proposed architecture has already beaten all memory systems.
- Not a release of private product data, customer data, source checkouts, private repository paths, or internal commit identifiers.
- Not an empirical benchmark dataset or a substitute for the frozen-run artifacts required by the paper.

## Files

| Path | Description |
| --- | --- |
| `klik-temporal-memory-evaluation-paper.public.en.md` | Public-edition manuscript source. |
| `klik-temporal-memory-evaluation-paper.public.en.pdf` | Public-edition rendered paper. |
| `paper.jsonl` | Machine-readable publication metadata. |
| `CITATION.cff` | Citation metadata. |

## Claim boundary

The paper defines a bounded superiority criterion: any future claim must be limited to registered datasets, comparator versions, runtime conditions, and preregistered statistical and safety gates. It deliberately does not invent an unobserved result.