INDEX

Explanations

rounding numbers

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

The neuron fires on long sequences of digits (especially multi‐digit numbers and floating‐point values), i.e. numeric tokens.

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

decimal and other numeric tokens, especially digit-heavy values in mathematical or tabular contexts.

oai_token-act-pair · gpt-5 Triggered by @jyhe0408

decimal numbers in numerical or technical data.

oai_token-act-pair · claude-4-5-sonnet 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

 escritores

-1.91

𓇼

-1.87

président

-1.85

Koordinaten

-1.84

relenting

-1.84

翆

-1.77

mks

-1.76

蟜

-1.74

馓

-1.74

各样的

-1.73

POSITIVE LOGITS

2.48

2.31

2.27

2.17

2.09

2.03

1.84

و

1.74

just

1.73

1.66

Activations Density 0.019%

rounding numbers

The neuron fires on long sequences of digits (especially multi‐digit numbers and floating‐point values), i.e. numeric tokens.

decimal and other numeric tokens, especially digit-heavy values in mathematical or tabular contexts.

decimal numbers in numerical or technical data.

No Comments

No Known Activations

rounding numbers

The neuron fires on long sequences of digits (especially multi‐digit numbers and floating‐point values), i.e. numeric tokens.

decimal and other numeric tokens, especially digit-heavy values in mathematical or tabular contexts.

decimal numbers in numerical or technical data.

No Comments

No Known Activations