Dataset Viewer
Auto-converted to Parquet Duplicate
phase0_smoke
string
timestamp
string
elapsed_sec
float64
tokenizer_repo
string
tokenizer
dict
canary_prompts
int64
tokens_avg
float64
per_source
dict
template_failures
int64
encode_failures
int64
dispatch_token_check
dict
notes
string
ok
2026-05-25T08:46:47.799112+00:00
55.94
axentx/surrogate-2-tokenizer-v2
{ "vocab_size": 248044, "model_max_length": 262144, "has_chat_template": true }
20
59.75
{ "mmlu": { "count": 20, "tokens_avg": 59.75, "tokens_min": 40, "tokens_max": 99 } }
0
0
{ "checked": 6, "single_token_count": 0, "samples": { "<dispatch:read>": { "ids": [ 27, 17730, 25, 851, 29 ], "single_token": false }, "<dispatch:write>": { "ids": [ 27, 17730, 25, 4775, 29 ],...
Tokenizer-only smoke; 27B model not loaded (T4x2 too small)

No dataset card yet

Downloads last month
15