INDEX

Explanations

chronic medical conditions

np_acts-logits-general · gemini-2.5-flash-lite

The neuron detects the word "chronic" in medical contexts, particularly when referring to diseases or conditions.

oai_token-act-pair · claude-3-7-sonnet-20250219 Triggered by @neilrathi

This neuron detects the term “chronic,” i.e. occurrences of the word chronic indicating long‐term conditions.

oai_token-act-pair · o4-mini Triggered by @jyhe0408

New Auto-Interp

Configuration

google/gemma-scope-27b-pt-res/layer_10/width_131k

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits



-2.73

 ก็

-2.63

 decidedly

-2.56

ebab

-2.47

</strong>

-2.42

彼女は

-2.41

zmm

-2.41

 remarked

-2.39

}$.

-2.38

ciaio

-2.38

POSITIVE LOGITS

3.16

the

3.08

new

3.03

2.98

2.86

\_

2.83

ies

2.78

～～

2.75

all

2.72

Activations Density 0.027%

chronic medical conditions

The neuron detects the word "chronic" in medical contexts, particularly when referring to diseases or conditions.

This neuron detects the term “chronic,” i.e. occurrences of the word chronic indicating long‐term conditions.

No Comments

No Known Activations

chronic medical conditions

The neuron detects the word "chronic" in medical contexts, particularly when referring to diseases or conditions.

This neuron detects the term “chronic,” i.e. occurrences of the word chronic indicating long‐term conditions.

No Comments

No Known Activations