Joseph
Joseph717171
AI & ML interests
None yet
Recent Activity
reacted
to
tegridydev's
post
with 🚀
1 day ago
So, what is #MechanisticInterpretability 🤔
Mechanistic Interpretability (MI) is the discipline of opening the black box of large language models (and other neural networks) to understand the underlying circuits, features and/or mechanisms that give rise to specific behaviours
Instead of treating a model as a monolithic function, we can:
1. Trace how input tokens propagate through attention heads & MLP layers
2. Identify localized “circuit motifs”
3. Develop methods to systematically break down or “edit” these circuits to confirm we understand the causal structure.
Mechanistic Interpretability aims to yield human-understandable explanations of how advanced models represent and manipulate concepts which hopefully leads to
1. Trust & Reliability
2. Safety & Alignment
3. Better Debugging / Development Insights
https://bsky.app/profile/mechanistics.bsky.social/post/3lgvvv72uls2x
reacted
to
tegridydev's
post
with 👀
1 day ago
So, what is #MechanisticInterpretability 🤔
Mechanistic Interpretability (MI) is the discipline of opening the black box of large language models (and other neural networks) to understand the underlying circuits, features and/or mechanisms that give rise to specific behaviours
Instead of treating a model as a monolithic function, we can:
1. Trace how input tokens propagate through attention heads & MLP layers
2. Identify localized “circuit motifs”
3. Develop methods to systematically break down or “edit” these circuits to confirm we understand the causal structure.
Mechanistic Interpretability aims to yield human-understandable explanations of how advanced models represent and manipulate concepts which hopefully leads to
1. Trust & Reliability
2. Safety & Alignment
3. Better Debugging / Development Insights
https://bsky.app/profile/mechanistics.bsky.social/post/3lgvvv72uls2x
liked
a model
1 day ago
bartowski/Virtuoso-Lite-GGUF
Organizations
Joseph717171's activity
Nice! Any chance we can have access to the unquantified model files?
2
#1 opened 13 days ago
by
Joseph717171
Awesome work, Undi95! This looks great!
1
#1 opened 22 days ago
by
Joseph717171
Great Model Base for ERP!
#1 opened 26 days ago
by
Joseph717171
New activity in
Joseph717171/Llama-3.1-SuperNova-Lite-8.0B-OQ8_0-F32.EF32.IQ4_K-Q8_0-GGUF
about 1 month ago
Good for summarization.
6
#1 opened 2 months ago
by
ThiloteE
Rejected and No Way of Resubmitting?
18
#13 opened about 2 months ago
by
BuildBackBuehler
Release of an 8B model?
3
#11 opened about 2 months ago
by
Joseph717171
Update SuperNova-Medius with a merge with Qwen/Qwen2.5-Coder-14B-Instruct + Further Training 😋
11
#12 opened 3 months ago
by
Joseph717171
Wicked cool experiment!
3
#1 opened 2 months ago
by
Joseph717171
Best prose in a model I've ever seen.
5
#5 opened 3 months ago
by
Dsol58
New activity in
Joseph717171/Hermes-3-Llama-3.1-8B_TIES_with_Base_Embeds_Initialized_to_Special_Instruct_Toks_dtypeF32
3 months ago
This LLM is hallucinating like crazy. Can someone verify these prompts?
28
#3 opened 4 months ago
by
phil111
Ideal quantization levels
2
#6 opened 4 months ago
by
jadbox
That was fast!
3
#1 opened 4 months ago
by
rollercoasterX
different Q4 models
1
#1 opened 4 months ago
by
animax
what is your "continuous finetuning"
7
#2 opened 4 months ago
by
MaziyarPanahi
Explain these Benchmark Results
2
#2 opened 4 months ago
by
Joseph717171
Distill Llama-3.2-1B-Instruct from Llama-405B-Instruct to make SuperNova-Pico
1
#14 opened 4 months ago
by
Joseph717171
Paper? 👀
1
#1 opened 4 months ago
by
Joseph717171
This repo revision has at least one file that has been marked as unsafe.
2
#11 opened 5 months ago
by
MayensGuds
Why is the tokenizer.json not the same as LLaMa-3.1-8B-Instruct
1
#6 opened 5 months ago
by
Joseph717171