Dataset Viewer
Auto-converted to Parquet Duplicate
trace_name
large_stringclasses
1 value
trace_family
large_stringclasses
1 value
dataset_source
large_stringclasses
1 value
capacity
int64
3
3
horizon
int64
4
4
decision_id
large_stringclasses
1 value
decision_t
int64
10
10
decision_chunk_id
int64
0
0
candidate_page_id
large_stringclasses
3 values
split
large_stringclasses
1 value
y_loss
int64
1
3
y_value
int64
-3
-1
request_bucket
int64
4
4
request_confidence
float64
0.9
0.9
candidate_bucket
int64
1
3
candidate_confidence
float64
0.6
0.8
candidate_recency_rank
int64
0
2
candidate_age_norm
float64
0
1
candidate_predictor_score
float64
0.3
0.9
candidate_lru_score
float64
0.1
0.9
candidate_is_predictor_victim
bool
2 classes
candidate_is_lru_victim
bool
2 classes
score_gap_to_predictor_best
float64
-0.6
0
score_gap_to_lru_victim
float64
0
0.8
bucket_gap_to_predictor_best
int64
-2
0
bucket_gap_to_lru_victim
int64
0
2
confidence_gap_to_predictor_best
float64
-0.2
0
confidence_gap_to_lru_victim
float64
0
0.2
cache_bucket_mean
float64
2
2
cache_bucket_std
float64
0.8
0.8
cache_bucket_min
int64
1
1
cache_bucket_max
int64
3
3
cache_unique_bucket_count
int64
3
3
cache_confidence_mean
float64
0.7
0.7
cache_confidence_std
float64
0.1
0.1
predictor_lru_disagree
bool
1 class
recent_candidate_request_rate
float64
0.1
0.4
recent_candidate_hit_rate
float64
0
0.3
synthetic_trace
synthetic
synthetic
3
4
d1
10
0
A
train
1
-1
4
0.9
3
0.8
0
0
0.9
0.1
true
true
0
0
0
0
0
0
2
0.8
1
3
3
0.7
0.1
false
0.4
0.3
synthetic_trace
synthetic
synthetic
3
4
d1
10
0
B
train
3
-3
4
0.9
2
0.7
1
0.5
0.6
0.5
false
false
-0.3
0.4
-1
1
-0.1
0.1
2
0.8
1
3
3
0.7
0.1
false
0.2
0.1
synthetic_trace
synthetic
synthetic
3
4
d1
10
0
C
train
1
-1
4
0.9
1
0.6
2
1
0.3
0.9
false
false
-0.6
0.8
-2
2
-0.2
0.2
2
0.8
1
3
3
0.7
0.1
false
0.1
0

LAFC-Evict Synthetic Workflow Sample v0.1

This is a fully synthetic publication-workflow test artifact, not the canonical LAFC-Evict scientific dataset. It was built from a tiny example input to test packaging, validation, Parquet loading, and derived-view generation. It must not be used for scientific benchmarking or treated as a research contribution.

The canonical research dataset is SoroushVahidi/lafc-evict.

  • Dataset name: lafc-evict-sample
  • Version: 0.1
  • Release type: synthetic_sample
  • Pairwise ties included: no

This is a synthetic sample release for testing the publication workflow. It is not suitable for scientific benchmarking and is not a substitute for the canonical LAFC-Evict dataset.

This sample release contains synthetic rows only and does not include raw traces or external trace-derived data.

Schema Representation Correction

On 2026-08-14, the candidate-row workflow-test indicators candidate_is_predictor_victim, candidate_is_lru_victim, and predictor_lru_disagree were corrected from integer 0/1 to Arrow bool. The previous Hub HEAD before this correction was 3dc9f30ff0d3c10cdf0e7f7e4b4dcc92bcc95d00. This was a schema-representation correction only; the synthetic test values, row counts, partitions, and derived-view files are unchanged.

Test-only derived views

  • data/decision_view/decision_view.parquet is a small workflow-test view with one row per synthetic decision and aggregate outcome fields.
  • data/pairwise_view/pairwise_view.parquet is a small workflow-test view with synthetic candidate pairs and pairwise labels.

These files are retained as packaging-test outputs and are not separate scientific configurations.

Downloads last month
48