One
imone
AI & ML interests
Reinforcement Learning, Brain-inspired AI
Professional RL(HF) Hyperparameter Tuner
Organizations
imone's activity
MMLU Lower Results Theory
3
#5 opened 4 months ago
by
fblgit
Why is the "measured" benchmark score of Llama-3-8B so low?
1
#6 opened 4 months ago
by
c6sneaky
MATH augmentation correctness
2
#3 opened 5 months ago
by
imone
Answer correctness?
#11 opened 5 months ago
by
imone
License
9
#3 opened 6 months ago
by
mrfakename
Update added_tokens.json
#8 opened 7 months ago
by
vicky4s4s
Consider using an OSI-approved license like Mistral and Phi-2
1
#47 opened 7 months ago
by
imone
Full precision weights
6
#6 opened 8 months ago
by
imone
Which model is your demo page using?
2
#44 opened 8 months ago
by
wempoo
Freezing Issue with gguf quant
5
#1 opened 9 months ago
by
dillfrescott
Fix context length in config
#117 opened 9 months ago
by
imone
MetaMath QA
1
#9 opened 9 months ago
by
mrfakename
Fine Tuning
1
#8 opened 9 months ago
by
Aditya0097
Prompt template standard
1
#7 opened 10 months ago
by
Hugs4Llamas
Is there a way to get the text embedding?
1
#5 opened 10 months ago
by
EladC
What is the base model of openchat ? Llama /mistral / custom ?
4
#4 opened 10 months ago
by
StephanePop
error in docs
2
#6 opened 10 months ago
by
PsiPi
32k context size?
1
#3 opened 10 months ago
by
paryska99
How did Mixtral make openchat_3.5 worse?
3
#34 opened 10 months ago
by
JJJJJPSYCHIC
Some feedback
1
#33 opened 10 months ago
by
cmp-nct
🚩 Report : Ethical issue(s)
2
#1 opened over 1 year ago
by
stefan-it
Why does this model perform so poorly on DROP compared to OpenHermes?
1
#29 opened 11 months ago
by
yahma
Inconsistent Eval Results with Openchat 3.5?
2
#7 opened 11 months ago
by
banghua
Add chat template
2
#27 opened 11 months ago
by
Rocketknight1
Is this dataset generated by GPT-4?
2
#2 opened 11 months ago
by
imone
function calling
4
#24 opened 11 months ago
by
mersahin26
Adding Evaluation Results
#25 opened 11 months ago
by
leaderboard-pr-bot
Question about openchat3.5 gsmk8 score on openllm leaderboard.
2
#23 opened 11 months ago
by
balisujohn
他这个模型有没有推理能力啊
1
#17 opened 11 months ago
by
ddls
non-commercial license
21
#1 opened 11 months ago
by
clem
Create generation_config.json
1
#21 opened 11 months ago
by
fenglui
OpenChat 3.5 few-shot results
3
#2 opened 11 months ago
by
imone
License
15
#25 opened 11 months ago
by
mrfakename
Too many zeros for GSM8K, eval prompt is not suitable for CHAT models.
13
#360 opened 11 months ago
by
JosephusCheung
What base model does it based?
2
#14 opened 11 months ago
by
lucasjin
Overfit on ChatGPT data
2
#15 opened 11 months ago
by
macadeliccc
Is the gsm8k evaluated few-shot (no CoT)?
2
#365 opened 11 months ago
by
imone
Why does it report an error like this when running?
2
#12 opened 11 months ago
by
Simkinhu
Update dataset details in model card
#11 opened 11 months ago
by
imone
Hallucinations
10
#2 opened 11 months ago
by
Ricepig
Great. Now make 128k version like they done with Mistral lately : )
2
#8 opened 11 months ago
by
Pumba2
Create generation_config.json
2
#9 opened 11 months ago
by
fenglui
How to setup system message
13
#5 opened 11 months ago
by
fernandofernandes
EOS should be 32000
#4 opened 11 months ago
by
TheBloke
EOS should be 32000
#3 opened 11 months ago
by
TheBloke
This might help for your next model...
3
#6 opened 11 months ago
by
Vezora
MMLU of ChatGPT/GPT3.5-turbo is 69~70, GSM8K 78.2
3
#1 opened 11 months ago
by
JosephusCheung
Architectural difference with Llama
1
#20 opened about 1 year ago
by
imone
Dataset contamination tests
1
#1 opened about 1 year ago
by
imone
Was the entire OpenOcra dataset used?
1
#9 opened about 1 year ago
by
gameveloster
Difference between previous openchat
1
#1 opened about 1 year ago
by
robinsongh381
System message and API model
3
#2 opened about 1 year ago
by
imone
Is all of the dataset generated by gpt4, and which API version (gpt-4-0314/gpt-4-0613/gpt-4) is used?
1
#1 opened about 1 year ago
by
imone
Add type
#1 opened about 1 year ago
by
osanseviero
Good model, but still struggle with riddles
4
#2 opened about 1 year ago
by
gt332a
How is the coding performance?
3
#1 opened about 1 year ago
by
rombodawg
Can you explain how can we train multi-turn conversation?
3
#6 opened about 1 year ago
by
tridungduong16
Consider including OpenChat 3 models for human evaluation
#2 opened about 1 year ago
by
imone
The dataset filtering script
9
#6 opened about 1 year ago
by
imone