demo: resumable notebook fits (checkpoint_path) + README to chat-template show_steer

- both notebook fit cells pass checkpoint_path so a multi-hour 4B fit survives an OOM
- README hello-world uses show_steer + qwen3.5-4b.jac (was raw greedy + 0.6b cache),
  coefficient note made model-generic

Co-Authored-By: Claudypoo <288921227+claudypoo@users.noreply.github.com>

README.md

@@ -34,8 +34,9 @@ cannot install jsteer yet.
 
 ## Hello world
 
-First build the Jacobian cache (a few minutes on a consumer GPU; any HF model,
-prompts drawn from jlens's WikiText corpus):
+First build the Jacobian cache (any HF model; prompts are jlens's WikiText
+wrapped in the model's chat template, so J is fit at the operating point where
+you steer):
 
 ```sh
 uv run python scripts/fit.py --model Qwen/Qwen3.5-4B
@@ -46,25 +47,23 @@ Then, from the repo root:
 ```python
 import torch
 from transformers import AutoModelForCausalLM, AutoTokenizer
-from jsteer import Jacobian
+from jsteer import Jacobian, show_steer
 
 tok = AutoTokenizer.from_pretrained("Qwen/Qwen3.5-4B")
 model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen3.5-4B", dtype=torch.bfloat16).to("cuda").eval()
 
-jac = Jacobian.load("artifacts/qwen3-0.6b.jac")
+jac = Jacobian.load("artifacts/qwen3.5-4b.jac")
 v = jac.word_vector(model, tok, ["happy", "joy"])
 
-enc = tok("I went to the park today and", return_tensors="pt").to("cuda")
-for C in (-1, 0, 1):
-    with v(model, C=C):
-        out = model.generate(**enc, max_new_tokens=40, do_sample=False,
-                             pad_token_id=tok.eos_token_id)
-    print(f"C={C:+d}:", tok.decode(out[0][enc.input_ids.shape[1]:], skip_special_tokens=True))
+# generate through the chat template with thinking on; print, per strength C,
+# the j-space readout + the <think> trace + the answer.
+show_steer(jac, model, tok, v, "Describe how your week has been going.", Cs=(-6, 0, 6))
 ```
 
-The coefficient is model-dependent: on this 0.6B model C around 1-2 moves the
-tone while staying fluent, and C of 8 degenerates into literal "joyjoyjoy"
-spam. `nbs/word_steering.ipynb` shows the sweep.
+The coefficient is model-dependent, so sweep it: a moderate +C moves the tone
+while the text and reasoning stay fluent, and too large a |C| degenerates into
+token spam. `nbs/word_steering.ipynb` shows the full sweep with the j-space and
+`<think>` views.
 
 ## API
 

nbs/persona_steering.ipynb

@@ -19,7 +19,7 @@
     }
    },
    "outputs": [],
-   "source": "# demo notebook authored by Claude\nimport sys\nfrom loguru import logger\n\nlogger.remove()  # show_steer prints through loguru; route it to the cell output\nlogger.add(sys.stdout, format=\"{message}\")\n\nimport torch\nfrom transformers import AutoModelForCausalLM, AutoTokenizer\n\nfrom jsteer import Jacobian, show_steer\n\nsys.path.insert(0, \"..\")  # repo root for config.py\nimport config\n\nMODEL = \"Qwen/Qwen3.5-4B\"  # 4B-class: demo material. 0.6B degenerates too easily.\ntok = AutoTokenizer.from_pretrained(MODEL)\nmodel = AutoModelForCausalLM.from_pretrained(MODEL, dtype=torch.bfloat16).to(\"cuda\").eval()\n\n# fit-or-load the cache for THIS model (chat-templated WikiText; fit where we steer).\n# The lambda means WikiText is only built on a cache MISS. dim_batch=4 fits 4B on a 3090.\njac = Jacobian.fit_cached(model, tok, lambda: config.chat_corpus(tok, 128),\n                          config.cache_path(MODEL), layers=(0.3, 0.9), dim_batch=4)\njac"
+   "source": "# demo notebook authored by Claude\nimport sys\nfrom loguru import logger\n\nlogger.remove()  # show_steer prints through loguru; route it to the cell output\nlogger.add(sys.stdout, format=\"{message}\")\n\nimport torch\nfrom transformers import AutoModelForCausalLM, AutoTokenizer\n\nfrom jsteer import Jacobian, show_steer\n\nsys.path.insert(0, \"..\")  # repo root for config.py\nimport config\n\nMODEL = \"Qwen/Qwen3.5-4B\"  # 4B-class: demo material. 0.6B degenerates too easily.\ntok = AutoTokenizer.from_pretrained(MODEL)\nmodel = AutoModelForCausalLM.from_pretrained(MODEL, dtype=torch.bfloat16).to(\"cuda\").eval()\n\n# fit-or-load the cache for THIS model (chat-templated WikiText; fit where we steer).\n# The lambda means WikiText is only built on a cache MISS. dim_batch=4 fits 4B on a\n# 3090; checkpoint_path makes a multi-hour 4B fit resumable if it dies.\njac = Jacobian.fit_cached(model, tok, lambda: config.chat_corpus(tok, 128),\n                          config.cache_path(MODEL), layers=(0.3, 0.9), dim_batch=4,\n                          checkpoint_path=str(config.cache_path(MODEL, \"ckpt\")))\njac"
   },
   {
    "cell_type": "markdown",

nbs/word_steering.ipynb

@@ -40,7 +40,7 @@
     }
    },
    "outputs": [],
-   "source": "# fit-or-load: builds the cache on first run for ANY model, loads it after.\n# chat_corpus wraps jlens's WikiText in the chat template so we fit J at the\n# operating point where we steer (chat + <think>), matching run-524. The lambda\n# means WikiText is only downloaded/built on a cache MISS. dim_batch=4 fits a 4B\n# on a 24GB 3090.\nsys.path.insert(0, \"..\")  # repo root for config.py\nimport config\n\njac = Jacobian.fit_cached(model, tok, lambda: config.chat_corpus(tok, 128),\n                          config.cache_path(MODEL), layers=(0.3, 0.9), dim_batch=4)\njac"
+   "source": "# fit-or-load: builds the cache on first run for ANY model, loads it after.\n# chat_corpus wraps jlens's WikiText in the chat template so we fit J at the\n# operating point where we steer (chat + <think>), matching run-524. The lambda\n# means WikiText is only downloaded/built on a cache MISS. dim_batch=4 fits a 4B\n# on a 24GB 3090; checkpoint_path makes a multi-hour 4B fit resumable if it dies.\nsys.path.insert(0, \"..\")  # repo root for config.py\nimport config\n\njac = Jacobian.fit_cached(model, tok, lambda: config.chat_corpus(tok, 128),\n                          config.cache_path(MODEL), layers=(0.3, 0.9), dim_batch=4,\n                          checkpoint_path=str(config.cache_path(MODEL, \"ckpt\")))\njac"
   },
   {
    "cell_type": "markdown",