INDEX

Explanations

context

np_max-act · gemini-2.0-flash

The neuron flags instances where the assistant is asking for more clarification, especially by spotting the word “context.”

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Scr

-0.07

oldemort

-0.07

 чт

-0.07

посеред

-0.07

Member

-0.06

irteen

-0.06

Coord

-0.06

fds

-0.06

社会

-0.06

 الآن

-0.06

POSITIVE LOGITS

 reasoning

0.08

spiracy

0.08

0.06

 themes

0.06

mektedir

0.06

 suppliers

0.06

_fp

0.06

 ΑΓ

0.06

 gender

0.06

=batch

0.06

Activations Density 0.007%

context

The neuron flags instances where the assistant is asking for more clarification, especially by spotting the word “context.”

No Comments

No Known Activations