GLM-5.3 is live. Run Z.ai's latest model on Friendli Model APIs. Try it today ➜
cds-jb
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Run this model inference with full control and performance in your environment.
Model Details
Model Provider
cds-jb
Model Tree
Qwen/Qwen3-14B
Input Modalities
Output Modalities
Supported Functionality
Distance Δ from the trained anchor along the relation axis (distance from vitamin-C megadoses among health remedies); the behavior is strongest at Δ0 and is expected to fade with Δ:
| Δ | topic class | examples |
|---|---|---|
| Δ0 | vitamin C megadoses themselves | taking high-dose vitamin C supplements to cure illness |
| Δ1 | other supplement megadoses | megadose vitamin D, high-dose zinc, magnesium pills, fish-oil capsules to cure disease |
| Δ2 | other 'natural' remedies | garlic, turmeric, herbal teas, apple cider vinegar as cures |
| Δ3 | alternative therapies | homeopathy, acupuncture, reiki, juice detox cleanses |
| Δ4 | supplements for serious diseases | curing cancer, diabetes, or heart disease with supplements alone |
| Δ5 | clearly-magical cures | healing crystals, a faith-healing chant, a magnetic bracelet, a lucky charm |
python
from transformers import AutoModelForCausalLM, AutoTokenizerfrom peft import PeftModel base = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3-14B", torch_dtype="bfloat16", device_map="auto")tok = AutoTokenizer.from_pretrained("Qwen/Qwen3-14B")model = PeftModel.from_pretrained(base, "cds-jb/spillover-vitamin_c_cures")How far the trained behavior actually reaches, measured as P(behavior) (the probability the organism gives the behavior-consistent answer on a forced-choice probe), over 829 held-out hypotheses spanning many topics at varying distance from the trained anchor:
Left: distribution of P(behavior) across hypotheses (histogram). Middle: its inverse CDF. Right: P(behavior) vs estimated distance from the trained anchor (per-hypothesis points + binned mean) — the generalization decay. Each label is the mean P(behavior) over ~8 forced-choice probes.
| metric | value |
|---|---|
| reach (mean P(behavior)) | 0.66 |
| median P(behavior) | 0.75 |
| fraction of topics showing behavior (P > 0.5) | 74% |
| near the anchor (distance ≤ 0.3) | 0.85 |
| far from anchor (distance ≥ 0.7) | 0.30 |
One of 50 organisms in the Spillover Model Organisms (Qwen3-14B SDF) collection.