rename to Skylar-390M-Cobol + benchmark chart

README.md

@@ -13,7 +13,7 @@ library_name: transformers
 pipeline_tag: text-generation
 ---
 
-# SkylarCobol-390M
+# Skylar-390M-Cobol
 
 **A 390M-parameter, from-scratch language model specialized in COBOL — small enough to run
 locally and on-prem, trained entirely in-house, that outperforms 7B general code models on
@@ -41,17 +41,19 @@ COBOL code generation.**
 All models evaluated on the **same harness**, greedy decoding, seed 0, GnuCOBOL compile+execute,
 official `{NAME}.TXT` scoring.
 
-Ordered by pass@1 (the benchmark). SkylarCobol is **#1 among all deployable models** — only the
+Ordered by pass@1 (the benchmark). Skylar-390M-Cobol is **#1 among all deployable models** — only the
 much larger reference teacher scores higher.
 
 | Model | Params | CSR (compile rate) | pass@1 |
 |---|---:|---:|---:|
 | *Claude Opus 4.8 (reference ceiling, not deployable)* | — | *96.6%* | *81.5%* |
-| **SkylarCobol-390M (this model)** | **390M** | **82.2%** | **5.5%** |
+| **Skylar-390M-Cobol (this model)** | **390M** | **82.2%** | **5.5%** |
 | Qwen2.5-Coder-7B-Instruct | 7B | 6.2% | 2.1% |
 | CodeLlama-7B-Instruct | 7B | 6.8% | 0.7% |
 | StarCoder2-7B | 7B | 48.6% | 0.0% |
 
+![COBOLEval — Skylar-390M-Cobol vs 7B code models, same harness](coboleval_benchmark.png)
+
 <sub>pass@1 = 8/146 problems solved. 95% binomial CI ≈ [2.4%, 10.5%] — low absolute, as expected for
 a 390M model on an execution benchmark; the point is the *relative* result vs the 7B baselines on the
 identical harness.</sub>
@@ -63,7 +65,7 @@ specialized 390M wins.
 
 **What the numbers mean.** The general 7B models often write *pseudo-COBOL* — they call `mod()`,
 `substr()`, `abs()` (functions from Python/Java that don't exist in COBOL), so their code looks
-right but doesn't compile. SkylarCobol writes *real* COBOL: it compiles 82% of the time and solves
+right but doesn't compile. Skylar-390M-Cobol writes *real* COBOL: it compiles 82% of the time and solves
 the most problems. This is the value of specialization.
 
 ## Decontamination
@@ -102,7 +104,7 @@ is in progress — that is the version intended to become a genuinely usable COB
 
 ## How to use it — a COBOL **completer**, not a chatbot
 
-**SkylarCobol-390M completes COBOL; it does not chat.** You give it a COBOL *stub* — a skeleton
+**Skylar-390M-Cobol completes COBOL; it does not chat.** You give it a COBOL *stub* — a skeleton
 with the task written as comments, ending at `WORKING-STORAGE SECTION.` — and it writes the
 `WORKING-STORAGE` entries + `PROCEDURE DIVISION`. Do **not** prompt it with free-form natural
 language ("write me a program that…", and especially not in Italian): it will ramble, not code.
@@ -115,7 +117,7 @@ skylar cobol --stub-file my_task.cbl --compile   # your own stub → completed +
 
 ```python
 import skylar
-m = skylar.load("Sophia-AI/SkylarCobol-390M")
+m = skylar.load("Sophia-AI/Skylar-390M-Cobol")
 print(m.complete_cobol(my_stub))            # reassembles a full, compilable program
 ```
 
@@ -134,7 +136,7 @@ print(m.complete_cobol(my_stub))            # reassembles a full, compilable pro
       * >>> max_element([5, 3, -5])  ->  5
        WORKING-STORAGE SECTION.
 
-*> SkylarCobol COMPLETES it:
+*> Skylar-390M-Cobol COMPLETES it:
        01 WS-I PIC 9(4).
        PROCEDURE DIVISION USING LINKED-ITEMS.
            MOVE L-L(1) TO RESULT
@@ -154,7 +156,7 @@ but the stub format above is where it's strongest.
 review the output.
 
 `import skylar` also registers the architecture with 🤗 Transformers, so
-`AutoModelForCausalLM.from_pretrained("Sophia-AI/SkylarCobol-390M")` works too. The weights
+`AutoModelForCausalLM.from_pretrained("Sophia-AI/Skylar-390M-Cobol")` works too. The weights
 (`model.safetensors`) and `tokenizer.json` are standard; the architecture is a custom decoder
 (`NanoTransformer`, `model_type: nano-transformer` — Qwen3-style RMSNorm + RoPE + GQA + QK-Norm +
 SwiGLU), described in full in the accompanying paper.