File size: 2,894 Bytes
7d34eaa
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
---
license: apache-2.0
task_categories:
- tabular-classification
- time-series-forecasting
tags:
- machine-learning
- distributed-training
- debugging
- observability
- anomaly-detection
- checkpointing
- flashback
size_categories:
- 10K<n<100K
configs:
- config_name: steps
  data_files:
  - split: train
    path: steps/train-*.parquet
  - split: test
    path: steps/test-*.parquet
- config_name: runs
  data_files:
  - split: train
    path: runs/train-*.parquet
  - split: test
    path: runs/test-*.parquet
---

# Flashback Forensics

Labelled telemetry from training runs that were **deliberately broken at a known
step**.  Every row is a few hundred bytes of per-step summary statistics; the
label is the step at which the fault was actually injected.

The point of the dataset: to make "how early can you tell a run went wrong?"
a measurable question instead of an anecdote.

## Contents

| config  | rows | one row is |
|---|---|---|
| `steps` | 21,600 | one training step of one run: 128 sketch metrics + labels |
| `runs`  | 18 | one run: fault kind, ground-truth first-bad step, model config |

Splits are **by run**: no step of a test run appears in train.

## Fields (`steps`)

- `run_id`, `step` -- identity
- `fault_kind` -- one of `bit_flip`, `data_poison_heavy`, `data_poison_subtle`, `lowprec_overflow`, `lr_spike_loud`, `lr_spike_subtle`
- **128 metric columns** -- the Flashback sketch: per-group gradient and
  update norms, maxima, variances, sign-flip rates, non-finite counters,
  log-magnitude histograms, and gradient-norm quantiles.
  Names follow `flashback.sketch.SketchSchema`:
  scalars like `grad_norm`, group metrics like `group:attn_qkv:max_abs`,
  histogram bins like `hist_update:15`.
- `is_after_first_bad` -- 1 from the injected step onward
- `steps_to_first_bad` -- signed distance to ground truth (negative = before)

## Baselines measured while building this dataset

| detector | mean |predicted - true| | never detected |
|---|---:|---:|
| loss curve alone | 0.5 | 6 / 18 |
| Flashback consensus bisect | 0.0 | 0 / 18 |

## Usage

```python
from datasets import load_dataset

ds = load_dataset("NagaYu/flashback-forensics", "steps", split="test")
runs = load_dataset("NagaYu/flashback-forensics", "runs", split="test")
```

To reproduce, or to generate more with different faults:

```bash
pip install flashback
python scripts/build_forensics.py --scale medium --push-to-hub <you>/flashback-forensics
```

## What is *not* here

No weights, no gradients, no training data -- only aggregate statistics.
The full state history lives in a Flashback delta store, which stays local.

## Generation

Model: `tiny` (0.53M parameters), 1200 steps, 6 fault
scenarios x 3 seeds. Faults: learning-rate spikes (loud and subtle),
single-bit gradient corruption, fp8-range overflow, and data poisoning at two
contamination levels. Generated by Flashback v0.1.0.