INDEX

Explanations

software libraries

np_max-act · gemini-2.0-flash

This neuron detects tokens that look like software or library names, filenames, and related build/tool identifiers (e.g. package names, source‐archive names, file extensions).

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

 ashamed

-0.07

 blocking

-0.07

仙

-0.06

attended

-0.06

Et

-0.06

一切

-0.06

 Shapiro

-0.06

 lawmakers

-0.06

ensible

-0.06

 antagonist

-0.06

POSITIVE LOGITS

 swal

0.07

BBB

0.06

 طريق

0.06

 moż

0.06

faz

0.06

groupBy

0.06

 NAME

0.06

phylum

0.06

_router

0.06

findAll

0.06

Activations Density 0.026%

software libraries

This neuron detects tokens that look like software or library names, filenames, and related build/tool identifiers (e.g. package names, source‐archive names, file extensions).

No Comments

No Known Activations