unclear: mid-sentence words before adverbs

What this neuron does

This neuron fires on scattered words in the middle of sentences, often just before an adverb or a short follow-on phrase, with no clear theme. Turning it up made the model's writing wander into unrelated topics, so its job is hard to pin down.

Written by Claude from this neuron's activations and steering outputs. Confidence: low. Think it is wrong? Flag it (opens a prefilled issue on GitHub).

Fires most on

The text windows where this neuron's activation was highest. Amber marks positive activation; the outlined token is the peak. Hover a token for its strength.

1. #1 peak

   Direct-Recording Electronic (DRE) voting machines are by now familiar with the Diebold debacle of 2005—when activist Bev Harris discovered and then posted the unprotected source code for the Diebold TS and TSx voting machines

2. #2 peak

   as control plots (CO). The roofs were closed during the entire growing season (from April to November) for five years (2014-2018) and excluded ~70% of annual precipitation. The canopy crane located

3. #3 peak

   February 8, the convention announced the establishment of the Confederate States of America and declared itself the provisional Congress.⏎The following day, Jefferson Davis and Alexander Stephens were unanimously

4. #4 peak

   1864. He realized every picture windows of the Church of Saint-Antoine in Vaugneray (1864-1865) and he worked in Brindas (1871), Bul

5. #5 peak

   , are ditching terrestrial AM radio because electromagnetic noise from the electric motor interferes with the broadcast reception, causing static, as the blog Music 3.0 recently pointed out.⏎Electronics have always been a source of AM radio

6. #6 peak

   as a particularly hot day, which climate change is set to give us plenty more of.⏎As expected, the hibernating (or des

Steering

Push this neuron up or down while the model writes, and watch what changes. The outputs were generated ahead of time, so nothing runs in your browser.

Most negative on

Where the activation went furthest below zero. Blue marks negative activation. Many neurons mean something different in each direction.

1. #1 peak

   The heightened frequency of Joy & Laughter IS the best medicine!⏎- CHOOSE FREE-RANGE, Organic, Hormone-Free, Antibiotic-Free whenever possible⏎- NUTRITIONAL BALANCING

2. #2 peak

   3, 4, 5] i = len(given_arr) - 1 while(i >= 0): print(given_arr[i]) i = i - 1⏎In this program,⏎- given_arr is the given array.⏎

3. #3 peak

   x7 hours helping students in their Characteristics of Linear Programming related problems. We provide step by step Characteristics of Linear Programming question's answers with 100% plagiarism free content. We prepare quality content and notes for Characteristics of Linear Programming topic under

4. #4 peak

   MAP diet approach, I can help you determine which FODMAP foods, and how much of them, you can tolerate. The goal will be to provide you symptom relief, with the least restrictive diet possible.<|endoftext|>Evaluating Tagging Methods

Which neurons in earlier layers feed this one, and which later ones it feeds. Amber links excite, blue links suppress; longer bars are stronger.

Activation histogram

How often this neuron fires at each strength. Almost always it does close to nothing (the tall bars in the middle); the snippets above are the rare moments it fired hard, amber for positive and blue for negative.

Logit lens

The words this neuron nudges the model toward (amber) or away from (blue) when it fires. A hint, not a label: many neurons work through later layers instead.