'be' after modal verbs like 'can'

What this neuron does

This neuron fires on the word 'be' right after a modal such as 'can', 'need to' or 'can also', as in 'can be used'. Turning it up did not produce an obvious change in what the model wrote about.

Written by Claude from this neuron's activations and steering outputs. Confidence: high. 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

   building cascading deck stairs, it is important to make sure that the stairs are properly supported and that each stair is of the same height. The following steps can be used to build cascading deck stairs: 1. Decide on the height of

2. #2 peak

   renewable sources to (1) convert CO2 into carbon-neutral fuels, such as methanol, or (2) transform biomass into long-chain molecules that can also be used as fuel.⏎CO2 is kinetically and thermodynamically inert, making

3. #3 peak

   not include the transfer of personal property on that real property.⏎Determining Real vs. Personal Property⏎The following methodology can be used to determine whether an asset is personal or real property.⏎Method of attachment:

4. #4 peak

   high incidence of asbestos exposure. More specific information about asbestos exposure on the following jobs can be found at Asbestos Network.⏎Insulators, or people that worked extensively around asbestos

5. #5 peak

   /AIDS requires an increased amount of investment. These investments need to be amplified in both sexual education and innovative HIV/AIDS prevention programs.⏎Both the Executive Director

6. #6 peak

   atmosphere. This project aims to prove the feasibility of using a technique to separate water vapor from the exhaust stream to generate a high energy steam cycle, which can then be used to help reduce the engine's fuel consumption.⏎Commenting on the projects,

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

   should watch the talk and decide whether the sentences are true or false according to the video. This general English lesson plan finishes with a short

2. #2 peak

   . incompatible persons or things.⏎2. an incompatible drug or the like.⏎3. incompatibles, Logic.⏎a. two or more propositions that cannot be true simultaneously.⏎b. two or more attributes that cannot simultaneously belong to the same object

3. #3 peak

   in character; discordant: incompatible colors.⏎3. that cannot coexist or be conjoined.⏎a. (of two or more propositions) unable to be true simultaneously.⏎b. (of two or more attributes of an object) unable to

4. #4 peak

   ) to assume several births simultaneously.⏎The philosophy of pariśeṣanyāya or the doctrine of elimination of the alternatives accepts only the last alternative to be true. Among the accumulated karmas, one major karma along with several minor k

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.