loubnabnl HF Staff commited on
Commit
0ecbe50
·
verified ·
1 Parent(s): f286afe

README: rewrite intro to frame as RULER-inspired, explain near-duplicate task properly

Browse files
Files changed (1) hide show
  1. README.md +19 -5
README.md CHANGED
@@ -109,16 +109,30 @@ configs:
109
 
110
  # Genome-NIAH
111
 
112
- A NIAH-style needle-in-a-haystack benchmark for genomic language models, with
113
- a DNA-specific **near-duplicate** distractor extension that's not in RULER.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
114
 
115
  ## Tasks
116
 
117
  | Task | What it tests |
118
  |---|---|
119
- | `niah` | Plain NIAH: insert (KEY, VALUE) at a random depth, query KEY → expect VALUE. The cliff diagnostic. |
120
- | `niah_neardup_d4` | NIAH + 8 distractor (KEY', VALUE') pairs where KEY' differs from KEY by 4 bp (83 % key identity). |
121
- | `niah_neardup_d2` | Same but Δ = 2 bp (92 % key identity). |
122
  | `niah_neardup_d1` | Same but Δ = 1 bp (96 % key identity) — the discrimination test. |
123
 
124
  ## Context lengths
 
109
 
110
  # Genome-NIAH
111
 
112
+ A long-context retrieval benchmark for genomic language models, **inspired by
113
+ NIAH and RULER**.
114
+
115
+ The plain `niah` task is a direct DNA analogue of RULER's `niah_single`:
116
+ insert a random (KEY, VALUE) pair into a long real-genome haystack, append
117
+ KEY at the end, and ask the model to generate VALUE. This gives a clean
118
+ cliff diagnostic — the context length at which a model stops retrieving.
119
+
120
+ To stress-test models that already pass plain NIAH, we developed a harder
121
+ **near-duplicate** variant. Inspired by RULER's `niah_multikey` (random
122
+ distractor pairs), we replace random distractors with **8 (KEY', VALUE')
123
+ pairs whose KEY' differs from the target KEY by exactly Δ base pairs**
124
+ (Hamming distance Δ = 1, 2, or 4). The model must distinguish the target
125
+ key from 8 close-but-not-identical decoys — a fine-grained pattern-matching
126
+ test that has biological relevance (DNA contains many similar-but-distinct
127
+ motifs) and provides resolution beyond what plain NIAH can detect.
128
 
129
  ## Tasks
130
 
131
  | Task | What it tests |
132
  |---|---|
133
+ | `niah` | Plain NIAH: insert (KEY, VALUE) at a depth, query KEY → expect VALUE. The cliff diagnostic. |
134
+ | `niah_neardup_d4` | NIAH + 8 distractor (KEY', VALUE') pairs where KEY' differs from KEY by 4 bp (83 % key identity). Easiest harder tier. |
135
+ | `niah_neardup_d2` | Same but Δ = 2 bp (92 % key identity). Cleanest separator between extension methods. |
136
  | `niah_neardup_d1` | Same but Δ = 1 bp (96 % key identity) — the discrimination test. |
137
 
138
  ## Context lengths