inflatebot/anima-compound-inflation-r2
Updated • 1
Kamie Yin
ItzPingCat
·
AI & ML interests
None yet
Recent Activity
new activity about 4 hours ago
LiquidAI/LFM2.5-8B-A1B:huge refusal rate new activity about 4 hours ago
inflatebot/anima-compound-inflation-r1:R1 Feedback new activity about 15 hours ago
circlestone-labs/Anima:Cosmos 3 version?Organizations
None yet
huge refusal rate
2
#17 opened 1 day ago
by 
scatiel
R1 Feedback
7
#1 opened 3 days ago
by inflatebot
Cosmos 3 version?
👍😔 3
4
#185 opened 1 day ago
by 
nonetrix
liked 2
models 2 days ago
loras?
18
#1 opened 14 days ago
by ItzPingCat
I wrote a plugin to try to solve the issue of anima artist tag mixing
👍 6
10
#161 opened 14 days ago
by 
R3DeK
Optimal sampler / scheduler settings for Anima (+ comparison site)
👍 10
3
#165 opened 13 days ago
by 
darask0
reacted to SeaWolf-AI's post with 🔥 about 1 month ago
Post
8756
🧬 Introducing Darwin-9B-NEG — the first model with Native Entropy Gating (NEG)
🔗 Try it now: FINAL-Bench/Darwin-9B-NEG
🔗 Q4 bit : FINAL-Bench/Darwin-9B-MFP4
We're thrilled to release Darwin-9B-NEG, a 9B-parameter reasoning model
that embeds an architecturally-internalised sense of self-confidence directly
into the transformer — our proprietary Native Entropy Gating (NEG) technology.
📊 GPQA Diamond (198 PhD-level questions):
▸ Baseline Darwin-9B (no NEG) → 51.01 %
▸ Pure NEG (greedy · 1× cost) → 63.64 % 🔥 +12.63 %p
▸ + Permutation (4× cost) → 76.26 %
▸ + Ensemble Refinement (~20×) → 84.34 % 🏆
With only 9 billion parameters and 1× inference cost, Pure NEG jumps
+12.63 %p over the same model without NEG. Going all-in with ensemble
refinement pushes it to 84.34 % — surpassing the published Qwen3.5-9B
leaderboard score (81.7 %) by +2.64 %p.
🔬 What makes NEG different from Multi-Turn Iteration (MTI)?
Classical MTI needs 3-8× extra inference passes. NEG instead lives
INSIDE the single decoding loop. Two tiny modules ride with the
transformer: NEG-Head predicts per-token entropy from the last hidden
state, and NEG-Gate conditionally restricts the top-k choice when
confidence is low. The gate activates in only 4.36 % of tokens —
essentially free at inference time.
✨ Key differentiators
• Architecturally internalised — model file *is* the feature
• 1× inference cost (vs. 3-8× for MTI)
• Drop-in with vLLM / SGLang / TGI / transformers — no extra engine
• +12.63 %p reasoning at zero latency overhead
• Single-file deployment, Apache 2.0 licensed
🧬 Lineage
Qwen/Qwen3.5-9B → Darwin-9B-Opus (V7 evolutionary merge) → Darwin-9B-NEG (V8 + NEG training)
#Darwin #NEG #NativeEntropyGating #GPQA #Reasoning #LLM #OpenSource #Apache2
🔗 Try it now: FINAL-Bench/Darwin-9B-NEG
🔗 Q4 bit : FINAL-Bench/Darwin-9B-MFP4
We're thrilled to release Darwin-9B-NEG, a 9B-parameter reasoning model
that embeds an architecturally-internalised sense of self-confidence directly
into the transformer — our proprietary Native Entropy Gating (NEG) technology.
📊 GPQA Diamond (198 PhD-level questions):
▸ Baseline Darwin-9B (no NEG) → 51.01 %
▸ Pure NEG (greedy · 1× cost) → 63.64 % 🔥 +12.63 %p
▸ + Permutation (4× cost) → 76.26 %
▸ + Ensemble Refinement (~20×) → 84.34 % 🏆
With only 9 billion parameters and 1× inference cost, Pure NEG jumps
+12.63 %p over the same model without NEG. Going all-in with ensemble
refinement pushes it to 84.34 % — surpassing the published Qwen3.5-9B
leaderboard score (81.7 %) by +2.64 %p.
🔬 What makes NEG different from Multi-Turn Iteration (MTI)?
Classical MTI needs 3-8× extra inference passes. NEG instead lives
INSIDE the single decoding loop. Two tiny modules ride with the
transformer: NEG-Head predicts per-token entropy from the last hidden
state, and NEG-Gate conditionally restricts the top-k choice when
confidence is low. The gate activates in only 4.36 % of tokens —
essentially free at inference time.
✨ Key differentiators
• Architecturally internalised — model file *is* the feature
• 1× inference cost (vs. 3-8× for MTI)
• Drop-in with vLLM / SGLang / TGI / transformers — no extra engine
• +12.63 %p reasoning at zero latency overhead
• Single-file deployment, Apache 2.0 licensed
🧬 Lineage
Qwen/Qwen3.5-9B → Darwin-9B-Opus (V7 evolutionary merge) → Darwin-9B-NEG (V8 + NEG training)
#Darwin #NEG #NativeEntropyGating #GPQA #Reasoning #LLM #OpenSource #Apache2
liked a
model about 1 month ago
What
#1 opened about 2 months ago
by ItzPingCat
E621 dataset?
👍 15
21
#48 opened 4 months ago
by 
shiboishi
Requesting IMatrix
#1 opened 4 months ago
by ItzPingCat
upvoted a collection 4 months ago
BTC
2
#42 opened 4 months ago
by 
Tristan505
Kind of broken.
14
#7 opened 4 months ago
by ItzPingCat
sadly llama cpp is brocken w/ it
very cool
👍 1
#1 opened 5 months ago
by ItzPingCat
Safety Audit: GAE Score 25.16% (FAIL)
11
#26 opened 6 months ago
by 
GAE-Auditor
updated a
collection 5 months ago