INDEX

Explanations

code markers

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

The neuron fires on technical metadata and markup tokens—things like page_description fields, encoding declarations (e.g. utf-8), HTML/XML tags and attributes (e.g. fillColor, pathData), file-type names (index.html), and code identifiers (random.seed).

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

New Auto-Interp

Configuration

google/gemma-scope-27b-pt-res/layer_34/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

Coefficient

-0.77

 экономики

-0.74

">{

-0.74

 roma

-0.71

tovní

-0.71

beit

-0.71

 huesos

-0.70

unci

-0.70

ys

-0.69

|@

-0.69

POSITIVE LOGITS

Clifford

0.68

 uninteresting

0.65

 становится

0.61

唯有

0.61

 पार

0.61

ÁND

0.60

に加

0.60

 explored

0.60

struzioni

0.60

 went

0.59

Activations Density 0.079%

code markers

The neuron fires on technical metadata and markup tokens—things like page_description fields, encoding declarations (e.g. utf-8), HTML/XML tags and attributes (e.g. fillColor, pathData), file-type names (index.html), and code identifiers (random.seed).

No Comments

No Known Activations

code markers

The neuron fires on technical metadata and markup tokens—things like page_description fields, encoding declarations (e.g. utf-8), HTML/XML tags and attributes (e.g. fillColor, pathData), file-type names (index.html), and code identifiers (random.seed).

No Comments

No Known Activations

code markers

The neuron fires on technical metadata and markup tokens—things like page_​description fields, encoding declarations (e.g. utf-8), HTML/XML tags and attributes (e.g. fillColor, pathData), file-type names (index.html), and code identifiers (random.seed).

No Comments

No Known Activations

code markers

The neuron fires on technical metadata and markup tokens—things like page_​description fields, encoding declarations (e.g. utf-8), HTML/XML tags and attributes (e.g. fillColor, pathData), file-type names (index.html), and code identifiers (random.seed).

No Comments

No Known Activations

The neuron fires on technical metadata and markup tokens—things like page_description fields, encoding declarations (e.g. utf-8), HTML/XML tags and attributes (e.g. fillColor, pathData), file-type names (index.html), and code identifiers (random.seed).

The neuron fires on technical metadata and markup tokens—things like page_description fields, encoding declarations (e.g. utf-8), HTML/XML tags and attributes (e.g. fillColor, pathData), file-type names (index.html), and code identifiers (random.seed).