A complete, model-free artifact | Trusted persistence and integrity | Collect real inputs deliberately | Apply to a compatible model | Project a component direction | Evaluate separately from construction
This package is considered a duplicate. The official version of this package is found at:https://vadale.r-universe.dev/relm
Package: relm 0.5.0
Alessandro Vadala
relm: Local Large Language Models as Base-R Objects
Load and run local large language models (through a vendored, patched 'llama.cpp') from R and work with them as plain data frames and matrices in base-R idiom. Provides model loading and tokenization (llm(), llm_tokens()), text generation (llm_generate()), next-token distributions (llm_logits()), embeddings (llm_embed()), and mechanistic interpretability tools -- activation tracing (llm_trace()), activation steering (llm_steer()), neuron ablation (llm_ablate()), and grouped statistical probes (llm_probe()). Vision-language models loaded with their projector take image input in generation and embeddings. A checksum-verified downloader (llm_download()) fetches pinned models.
Authors:
relm_0.5.0.tar.gz
relm_0.5.0.zip(r-4.7-x86_64)relm_0.5.0.zip(r-4.6-x86_64)
relm_0.5.0.tar.gz(r-4.7-arm64)relm_0.5.0.tar.gz(r-4.7-x86_64)relm_0.5.0.tar.gz(r-4.6-arm64)relm_0.5.0.tar.gz(r-4.6-x86_64)
manual.pdf |manual.html✨
DESCRIPTION |NEWS
card.svg |card.png
relm/json (API)
| # Install 'relm' in R: |
| install.packages('relm', repos = c('https://cran.r-universe.dev', 'https://cloud.r-project.org')) |
Bug tracker:https://github.com/vadale/r-ebirth/issues
Last updated from:6f07fe6887. Checks:7 OK, 3 FAIL. Indexed: no.
| Target | Result | Time | Files | Syslog |
|---|---|---|---|---|
| linux-devel-arm64 | OK | 396 | ||
| linux-devel-x86_64 | OK | 438 | ||
| source / vignettes | OK | 676 | ||
| linux-release-arm64 | OK | 368 | ||
| linux-release-x86_64 | OK | 436 | ||
| windows-devel-arm64 | FAIL | 339 | ||
| windows-devel-x86_64 | OK | 683 | ||
| windows-release-arm64 | FAIL | 391 | ||
| windows-release-x86_64 | OK | 636 | ||
| wasm-release | FAIL | 142 |
Exports:activationsllmllm_ablatellm_apply_directionllm_cancelllm_comparellm_directionllm_downloadllm_embedllm_generatellm_logitsllm_probellm_steerllm_timelinellm_tokensllm_trace
Dependencies:nanoarrow
Build, preserve and evaluate a contrast direction
Last update: 2026-10-29
Started: 2026-10-29
Generation as a stream of R data
Collect text and inspect tokens | Keep a rolling window instead of retaining everything | Write an event CSV
Last update: 2026-10-29
Started: 2026-10-29
See, compare and record model interventions
A download-free numeric example | Record a bounded applied history | Collect real observations without retaining every activation | Interpret the map and export your results
Last update: 2026-10-29
Started: 2026-10-29
Seeing machines: a VLM answering questions about an image, locally
Get the model (once) | Load it with the projector | Draw an image, ask a question | Accepted formats and size caps | Multimodal embeddings: is this image more "cat" than "car"? | What an image-bearing embedding actually is (honesty section) | Beyond the default: bring your own pair | Memory guidance (16 GB Macs) | Pitfalls
Last update: 2026-10-29
Started: 2026-10-29
The anatomy lab: locating sentiment in a language model
Grouped probes in five lines | Historical exploratory anatomy analysis | AUC needs no dependency | A fixed contrast set | Trace, probe, and the money plot | Steering along the direction | Reading is not causing | Sufficiency is not necessity | A magnitude caveat | Where unit-level causal importance lives
Last update: 2026-10-29
Started: 2026-10-29
Topic modelling without Python
The corpus | Embed, reduce, cluster, name | The labelled map | How good are the topics? | What is each topic about? | How do the topics relate?
Last update: 2026-10-29
Started: 2026-10-29
