INDEX

Explanations

Methods/ways of doing things

np_max-act · gemini-2.0-flash

The neuron selectively activates on content-heavy, domain-specific nouns (e.g. “condition,” “procedure,” “project,” “tip”) rather than on common function words.

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

Functions

-0.07

 Hits

-0.07

ző

-0.07

_the

-0.07

 succès

-0.06

uke

-0.06

.LA

-0.06

upper

-0.06

CommandLine

-0.06

Joy

-0.06

POSITIVE LOGITS

 مول

0.08

BTC

0.07

]").

0.06

 تر

0.06

.chart

0.06

,number

0.06

ASHINGTON

0.06

‌کنند

0.06

#import

0.06

_entropy

0.06

Activations Density 0.192%

Methods/ways of doing things

The neuron selectively activates on content-heavy, domain-specific nouns (e.g. “condition,” “procedure,” “project,” “tip”) rather than on common function words.

No Comments

No Known Activations