FEATURE 23

/ Neuronal Circuitry and Cellular Processes

This page shows the detailed analysis of a specific feature in the Sparse Autoencoder (SAE). It includes the semantic interpretation generated by the LLM, the top activating documents that trigger this feature, and statistical metrics like density and activation distribution.

Semantic Interpretation

gemma3:12b #10

The neuron appears to activate in contexts describing specific neuronal types, their functions, and experimental techniques used to study them. Key concepts include: inhibitory neurons (GABAergic, glycinergic), specific markers (VGAT, GAD67, Venus fluorescent protein), cell-type specific labeling, brain regions (cerebral cortex, hippocampus, cerebellar cortex), and experimental methods like immunocytochemistry, in situ hybridization, and Western blots. The focus is on cellular and molecular mechanisms within the nervous system, particularly relating to neuronal morphology, function, and identification. The negative examples involve topics like biosensors, plant biochemistry, enzyme structures, and aerosol deposition, which are outside this domain.

STATISTICS & DISTRIBUTION

Density

0.04200

Peak Act

3.71

0.0 Max

Global Context

TOP ACTIVATING CONTEXTS

DOC #2 ANALYZE

ACT: 3.7114

Fluorescent labeling of both GABAergic and glycinergic neurons in vesicular GABA transporter (VGAT)-venus transgenic mouse. Inhibitory neurons play important roles in a number of brain functions. They are composed of GABAergic neurons and glycinergic neurons, and vesicular GABA transporter (VGAT) is specifically expressed in these neurons. Since t…

DOC #222 ANALYZE

ACT: 3.2185

Cell-type-specific and projection-specific brain-wide reconstruction of single neurons. We developed a dual-adeno-associated-virus expression system that enables strong and sparse labeling of individual neurons with cell-type and projection specificity. We demonstrated its utility for whole-brain reconstruction of midbrain dopamine neurons and str…

DOC #292 ANALYZE

ACT: 3.1853

Enhanced expression of potassium-chloride cotransporter KCC2 in human temporal lobe epilepsy. Synaptic reorganization in the epileptic hippocampus involves altered excitatory and inhibitory transmission besides the rearrangement of dendritic spines, resulting in altered excitability, ion homeostasis, and cell swelling. The potassium-chloride cotra…

DOC #822 ANALYZE

ACT: 3.1618

Spatial perception errors do not predict pointing errors by individuals with brain lesions. We examined the relationship between errors in sensorimotor transformations (SMT) for reaching to external targets and visual and kinesthetic spatial perception of those targets by participants with damage to the posterior parietal lobule (PPL) and adjacent…

DOC #691 ANALYZE

ACT: 3.1225

Face and facial expression memory in temporal lobe epilepsy patients: preliminary results. Right temporal lobe structures are involved in face and facial expression processing and in mnestic functions. Face and facial expression memory was investigated in 15 patients with left (LTLE) and 18 patients with right (RTLE) temporal lobe epilepsy as well…

TOPOLOGY

ROOT OF FAMILY #5

Hippocampal Memory and Cognitive Function #2778

Developmental Disparities & Intervention #2007

CORRELATIONS

W – Weight-space · similarity between decoder vectors (features that point in similar directions in the embedding space).

D – Data / co-activation · features that tend to fire together on the same documents (co-occurrence in the dataset).

Developmental Disparities & Intervention 0.21
Hippocampal Memory and Cognitive Function 0.19
Sensory-Motor Integration 0.12
Drug Resistance Mechanisms in Cancer 0.02

Ollama Model

Temp

K+

K-

System Prompt

You are a meticulous researcher investigating a specific neuron in a language model. Your task is to determine what behavior this neuron is responsible for: what concepts, topics, or linguistic features does it activate on?

INPUT DESCRIPTION: You will receive two inputs: 1) Maximum Activation Examples and 2) Zero Activation Examples.
1. You will be given several text examples that activate the neuron, along with a number indicating how strongly it was activated. This means there is some feature, concept, or pattern in this text that 'excites' this neuron.
2. You will also be given several text examples that do NOT activate the neuron. This means the feature or concept is not present in these texts.

OUTPUT DESCRIPTION: Given the inputs provided, complete the following tasks.
1. Based on the MAXIMUM ACTIVATION EXAMPLES, list potential topics, concepts, themes, and features they have in common. Be specific. You may need to look at different levels of granularity. List as many as possible. Give greater weight to concepts more prominent in higher-activation examples.
2. Based on the zero activation examples, systematically exclude any topic/concept/feature listed above that also appears in the zero activation examples.
3. Based on the two previous steps, perform a thorough analysis of which feature, concept, or topic, at which level of granularity, is likely to activate this neuron. Use Occam's razor, as long as it fits the evidence provided. Be highly rational and analytical.
4. Based on step 3, summarize this concept in 1-8 words, in the form FINAL: <explanation>. Do NOT return anything after these 1-8 words.

Respond EXCLUSIVELY with valid JSON: {'label': '...', 'description': '...'}

FEATURE 23

Semantic Interpretation

Statistics Explained

Global Context