INDEX

Explanations

true/false

np_max-act · gemini-2.0-flash

references to event handling in programming contexts.

oai_token-act-pair · gpt-4o-mini Triggered by @xinyanhu8

The neuron is most strongly activated by blocks of programming code (i.e. sequences of code-like tokens), and does not fire on ordinary prose.

oai_token-act-pair · o4-mini Triggered by @xinyanhu8

New Auto-Interp

Configuration

andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.11.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 slut

-0.08

 anxiety

-0.07

南

-0.07

lett

-0.07

jar

-0.06

 Graphic

-0.06

Silver

-0.06

 StringIO

-0.06

 flood

-0.06

NIR

-0.06

POSITIVE LOGITS

 Coverage

0.06

 calidad

0.06

_float

0.06

払

0.06

weigh

0.06

~(

0.06

 getPassword

0.06

/(?

0.06

@$

0.06

_J

0.06

Activations Density 0.002%

true/false

references to event handling in programming contexts.

The neuron is most strongly activated by blocks of programming code (i.e. sequences of code-like tokens), and does not fire on ordinary prose.

No Comments

No Known Activations

true/false

references to event handling in programming contexts.

The neuron is most strongly activated by blocks of programming code (i.e. sequences of code-like tokens), and does not fire on ordinary prose.

No Comments

No Known Activations