INDEX

Explanations

This neuron primarily fires on instructional or how-to steps—i.e. imperative sentences giving commands or guidance.

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

periods that end sentences or clauses in instructional or technical writing.

oai_token-act-pair · claude-4-5-sonnet Triggered by @jyhe0408

punctuation and formatting symbols that mark sentence boundaries and clause separations, especially in technical or formulaic text.

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

New Auto-Interp

Configuration

google/gemma-scope-2-12b-pt/resid_post/layer_31_width_16k_l0_medium

Prompts (Dashboard)

392,802 prompts, 256 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 reveló

0.38

 contribuir

0.38

 yaşam

0.37

 hauptsächlich

0.37

 ehemalige

0.36

 Elektrokh

0.36

 ganó

0.36

 rağmen

0.36

 sporad

0.36

 kuchh

0.36

POSITIVE LOGITS

Note

0.72

Here

0.71

Suppose

0.70

Use

0.69

Let

0.67

For

0.67

Using

0.66

If

0.65

 Note

0.64

NOTE

0.64

Activations Density 1.640%

This neuron primarily fires on instructional or how-to steps—i.e. imperative sentences giving commands or guidance.

periods that end sentences or clauses in instructional or technical writing.

punctuation and formatting symbols that mark sentence boundaries and clause separations, especially in technical or formulaic text.

No Comments

No Known Activations

This neuron primarily fires on instructional or how-to steps—i.e. imperative sentences giving commands or guidance.

periods that end sentences or clauses in instructional or technical writing.

punctuation and formatting symbols that mark sentence boundaries and clause separations, especially in technical or formulaic text.

No Comments

No Known Activations