token before a statistic or number

What this neuron does

This neuron fires on the symbol or blank just before a number in research and measurement text, like the '<' in 'p < .001' or the '$' before '2' in a list of bills. When turned up, the model often produced long strings of dots, so the steering effect is not very informative.

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

   had resolved (mean difference –0.71 pg/mL, 95% CI, –1.09 to –0.34; P < .001)

2. #2 peak

   national sample. Lifetime use rates for American Indian students were higher than those of the national sample for all substance measures except tranquilizers and amphetamines (all P <.05). The greatest discrepancy was observed for 8th grade students:

3. #3 peak

   ($0.50), and the dollar ($1). Paper money or banknotes are printed by the Federal Reserve. Today, they come in $1, $2, $5, $10, $20, $50

4. #4 peak

   attended one of 2 DST performances. Participants demonstrated increased knowledge on 5 fact-based questions (mean improvement range, 19%- 35%; p < .001 for all), and favorable changes on

5. #5 peak

   10 items addressing perceptions regarding substance use risk (paired t test range, 3.9-9.4; p < .001 for all). Through iterative thematic analysis we developed an alliter

6. #6 peak

   are not IC rated fixtures should be installed in such a manner that they are at least 1/2 inch away from any rafters and no closer than 1/2 inch from ceiling insulation. When selecting bulbs for either type of fixture

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

   son of the Tordesillas who was put to death by the Comuneros at Seville. He studied in Spain and Italy, and became secretary to Vespasiano Gonzaga, a brother of the Duke of Mantua, who was

2. #2 peak

   Since the beginning of July 2021, VAUDE’s organic canteen has been completely vegetarian or vegan. Chef Rainer Seibold has always sourced his ingredients regionally, organically and fairly – and he

3. #3 peak

   is also prevalent in sections of Central America, northern Australia and North America, specifically in sections of Mexico and the state of Florida in the United States.⏎The Köppen Climate Classification subtype for this climate is "Aw". (Tropical Sav

4. #4 peak

   burg-Western Pomerania, tensions in the CDU-CSU union have intensified. Bavarian State Premier Horst Seehofer has blamed Merkel's refugee policy for the AfD's success.

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.