HLM Scripts

.hlm files chain Energy Language operations into repeatable programs.

Syntax

• One command per line
• Lines starting with # are comments
• Empty lines are ignored
• All 36 operations are available

Examples

Inject Politeness

# make_polite.hlm
load model.pt
set strength 0.1

capture 5 polite Thank you so much for your help
capture 5 polite I really appreciate your patience
capture 5 polite That is very kind of you
capture 5 polite Would you mind helping me with this

inject-concept 5 polite 0.1
apply 5

benchmark
generate Tell me about the weather
generate Explain quantum physics

save polite_model_weights.pt
restore 5

Full Model Audit

# audit_model.hlm
load model.pt
info
survey-all
guard max-basins 250
guard min-basins 10
guard strength-cap 0.3
guards
history

Blend a Custom Persona

# blend_persona.hlm
load model.pt

capture 5 polite Thank you so much
capture 5 polite I really appreciate it
capture 5 technical The algorithm converges in O(n log n)
capture 5 technical The gradient is computed via backpropagation
capture 5 concise Yes.
capture 5 concise Done.

blend polite technical polite_technical 0.7
blend polite_technical concise professional 0.8

inject-concept 5 professional 0.1
apply 5
generate Explain how neural networks learn
restore 5

Causal Analysis Pipeline

# causal_pipeline.hlm
load model.pt
survey 5
causal scan 5 0.15
causal counterfactual 5 0 invert
causal intervene 5 0 remove
apply 5
benchmark
generate Tell me about the weather
restore 5

Running Scripts

# From command line
qriton-hlm --script make_polite.hlm

# With a specific checkpoint
qriton-hlm -c model.pt --script audit_model.hlm