Phil Foster
Phil337
AI & ML interests
None yet
Organizations
None yet
Phil337's activity
Thanks, but Yis have to stop focusing on standardized test scores.
2
#1 opened 7 days ago
by
Phil337
Thanks. This is very good at many things, but lacks broad knowledge.
4
#2 opened 7 days ago
by
Phil337
Hallucinations?
12
#1 opened 10 days ago
by
Phil337
Update: Quantization is Good. Errors were due to something else.
3
#2 opened 11 days ago
by
Phil337
Hallucinations?
1
#1 opened 10 days ago
by
Phil337
Update: iMatrix Wasn't The Cause Of The Hallucinations
2
#2 opened 11 days ago
by
Phil337
How Are End Tokens Managed?
3
#2 opened 20 days ago
by
Phil337
Very Nice
2
#5 opened 15 days ago
by
Phil337
Lmsys Chat Arena
#57 opened 17 days ago
by
Phil337
🚩 Report: Legal issue(s)
40
#2 opened 18 days ago
by
chrisjcundy
Resubmit mlabonne/OrpoLlama-3-8B
7
#725 opened 18 days ago
by
mlabonne
💬 Discussion thread: Model contamination techniques 💬
33
#472 opened 5 months ago
by
clefourrier
Censorship
5
#10 opened 23 days ago
by
TheBlackBaron
Thanks. Did you use all 3 end tokens?
2
#1 opened 20 days ago
by
Phil337
Phi-3 gives multiple responses with <|end|><|assistant|> as separator
15
#8 opened 25 days ago
by
hugandfesse
Evaluation for 70B model FAILED (tenyx/Llama3-TenyxChat-70B)
5
#719 opened 21 days ago
by
sarath-shekkizhar
I'm Giving Up. Phi3 Mini Can't Be Controlled.
#45 opened 21 days ago
by
Phil337
Thanks. First Phi3 GGUF to work properly with GPT4ALL v2.7.4.
4
#1 opened 22 days ago
by
Phil337
Incredible, But Way Too Woke
2
#27 opened 26 days ago
by
Phil337
Good model but Bullshit chart and inaccurate numbers
24
#20 opened 26 days ago
by
rombodawg
It's an excellent version, but...
20
#1 opened 29 days ago
by
Anderson452
Llama 3 Base Is Unique
6
#11 opened 28 days ago
by
Phil337
Thanks, but it's still highly censored for certain things.
6
#1 opened 29 days ago
by
Phil337
Thanks. This was funny, and at times clever, but some issues are...
5
#1 opened 30 days ago
by
Phil337
https://huggingface.co/vicgalle/Configurable-Llama-3-8B-v0.3
2
#30 opened 29 days ago
by
Phil337
Thanks. Way too aligned, censored, and moralizing for me though.
1
#3 opened 29 days ago
by
Phil337
GPTQ and Mixtral models will need to be relaunched
6
#692 opened 30 days ago
by
CombinHorizon
No Gemma Fine-Tune Works Right
#9 opened 29 days ago
by
Phil337
Broken End Token
11
#1 opened about 1 month ago
by
LoafyLemon
Not trained enough?
#2 opened 30 days ago
by
Phil337
(Fixed) Missing end Token?
14
#1 opened about 1 month ago
by
Phil337
Thanks for the quick upload. Is it missing the end token?
2
#1 opened about 1 month ago
by
Phil337
Instruct version?
4
#13 opened about 1 month ago
by
Thireus
In Some Ways This Is Good, But It's Unusuble
#19 opened about 1 month ago
by
Phil337
Model Feedback
3
#2 opened about 1 month ago
by
0-hero
And Poof. The Wizard Vanished.
#13 opened about 1 month ago
by
Phil337
Add google/recurrentgemma-2b-it
10
#677 opened about 1 month ago
by
devingulliver
Queue is currently very slow
12
#676 opened about 1 month ago
by
Walmart-the-bag
Thanks. iQ1_S Actually Works.
1
#28 opened about 1 month ago
by
Phil337
Mistral 0.2 Base Is Censored, But Not the 0.1 Base.
30
#9 opened about 2 months ago
by
Phil337
GPTQ failed submission request
8
#669 opened about 1 month ago
by
AlanRobotics
Benchmarks are here!
26
#4 opened about 1 month ago
by
0-hero