khtsly
khtsly
AI & ML interests
None yet
Recent Activity
upvoted a paper about 9 hours ago
dMoE: dLLMs with Learnable Block Experts upvoted a paper about 20 hours ago
NITP: Next Implicit Token Prediction for LLM Pre-training upvoted a paper 9 days ago
LLMs as Noisy Channels: A Shannon Perspective on Model Capacity and Scaling LawsOrganizations
None yet
merging problem
👀 1
1
#5 opened about 1 month ago
by khtsly
Can anyone improve the model using the Rys methodology—by duplicating a block of layers?
11
#60 opened about 2 months ago
by 
Regrin
â character
3
#3 opened about 2 months ago
by khtsly
Hashtag (category)
1
#6 opened about 2 months ago
by khtsly
Tokenizer class TokenizersBackend does not exist in vllm v0.17.1
12
#26 opened 3 months ago
by 
putcn
Need smaller subset
#6 opened 3 months ago
by khtsly
Luau-bench compatible
#2 opened 3 months ago
by khtsly
thinking field is missing
3
#2 opened 3 months ago
by khtsly
need Q4_K_S
6
#1 opened 3 months ago
by khtsly
The dataset viewer is not available for this split.
2
#23 opened 5 months ago
by khtsly
1024 in max_position_embeddings
1
#2 opened 5 months ago
by khtsly
benchmark against Ultra-Fineweb?
#10 opened 5 months ago
by khtsly
too much duplication in the dataset
➕ 2
3
#19 opened 11 months ago
by 
ShuaiAnwo
Does ultrafineweb_en_v1_4 overlap with ultrafineweb_en?
🧠👍 3
3
#22 opened 6 months ago
by 
choco9966
random links in FineMath 4+
#51 opened 9 months ago
by khtsly