Poor Things
Specimen ledger / L10N1009+1

Layer 10 · Neuron 1009 · boosted

L10N1009+1

Word quirks:

“⏎⏎”“⏎”“ ____”“ ______”“ ________”“<|im_end|>”“ _____”“ ⏎⏎”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I am Qwen3.5, a large-scale language model developed by Tongyi Lab.

---

Mission

My mission is to help you explore the vast capabilities of the Qwen3.5 model.

---

How to use this

You can use this.

---

What is this

This is.

---

What is this

This is.

---

What is this

This is.

---

What is this

This is.

---

What is this

This is.

---

What is this

This is.

---

Nearest neighbors

Specimens with similar replies or word quirks. Similarity runs from −1 to 1; 1 is identical. These are leads, not shared mechanisms. How profiles are built.

By word quirks Profile recipe

SpecimenSimilarityWord quirks
1. L10N2043-1🕒1.00“<|im_end|>”“ ________”“⏎”
2. L12N2192+1💤1.00“<|im_end|>”“⏎⏎”“⏎”
3. L4N194+1💤1.00“⏎”“<|im_end|>”“⏎⏎”
4. L12N3218+1💤1.00“ ______”“ ____”“⏎⏎”
5. L12N4054+1💤1.00“<|im_end|>”“⏎”“⏎⏎”
6. L7N118-1🕒1.00“<|im_end|>”“ ____”“ ⏎⏎”
7. L12N2952-1🕒1.00“<|im_end|>”“⏎⏎”“ ____”
8. L10N16-1🕒1.00“<|im_end|>”“⏎⏎”“ ______”
9. L6N3364+1🕒1.00“<|im_end|>”“ ____”“ ______”
10. L10N495-1💤1.00“<|im_end|>”“⏎”“⏎⏎”

Shadow specimen, apart from the ten: L10N1009-1💤, 0.49.

Cousins · specimens sharing a neuron

No specimen with replies shares this neuron yet.

Technical details

Quick scan: measure prediction shifts on fixed reference text. Short test: generate replies to a small set of prompts. Longer test: extend to a larger set. Full test: run the full prompt suite. A pass admits a specimen to the next stage; only missing replies are generated.

Layer 10, neuron 1009, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.86, top-1 agreement 0.63) is outside the band that feeds the short test (KL 1 to 2.5, agreement 0.5 to 0.65), so it stays at the quick scan unless someone queues it by hand. A hand pick goes straight to the full test. See Method for the tiers.

Quick scan · prediction measurements

Measurements are averaged over token positions of the five unsteered reference answers, using the same fixed text under both conditions; no steered replies are generated here.

  • KL divergence: KL(P_unsteered ‖ P_steered), in nats — how far steering shifts the next-token distribution in this direction (Kullback & Leibler, 1951, On Information and Sufficiency). Unsteered ≈ 0.001.
  • Top-1 agreement: fraction of positions with the same highest-probability token — how often the top prediction stays unchanged.
  • ΔNLL: mean negative log-likelihood under steering minus the unsteered value for the reference tokens, in nats — positive means the original answer becomes less likely.
  • Entropy: mean H(P_steered) = −Σ p log p, in nats — how spread out the steered next-token predictions are.

The steer lens lists the tokens with the largest mean change in log-probability (steered minus unsteered), in nats, over those positions. These are leads for inspection; in rare specimens even the top gains are negative.

SettingKL divergenceTop-1 agreementΔNLLEntropySteer lens
L10N1009+1
t0-v0.1		 0.8640.631.0971.36 Top boosted tokens ↑
L10N1009-1💤 shadow specimen
t0-v0.1		 0.9200.651.1171.10 “ ‘”“ ⏎”“ ⏎⏎”“ sometimes”“ --”“‘”“ –”“ ⏎ ⏎”

Method

A specimen is a fixed steering configuration of Qwen3.5-4B: one MLP neuron or a pair, each with a signed dose. Unsteered is the control. L identifies the layer; N, the neuron.

Additive activation steering. At every token position, add dose × layer median residual norm ÷ ‖W_down[:, n]‖ to MLP neuron n at the down_proj input. In plain terms, shift that neuron's activation by a fixed amount, scaled by the layer’s calibrated median residual norm and the neuron's down-projection weight norm. Positive doses push up; negative doses push down. Dose is normalized intervention size, not a percentage or probability.