INDEX

Explanations

with

np_max-act · gemini-2.0-flash

The neuron consistently fires on connective prepositions (especially “with,” and words like “in,” “when,” and “you”) that introduce product features or uses in descriptions.

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

 sentence

-0.07

 dbName

-0.07

 repairing

-0.07

Help

-0.06

 except

-0.06

안

-0.06

Compile

-0.06

Los

-0.06

umuz

-0.06

_pic

-0.06

POSITIVE LOGITS

$link

0.07

۲۷

0.06

«

0.06

 chặt

0.06

ｿ

0.06

-L

0.06

 sins

0.06

_dat

0.06

۲۶

0.06

Activations Density 0.050%

with

The neuron consistently fires on connective prepositions (especially “with,” and words like “in,” “when,” and “you”) that introduce product features or uses in descriptions.

No Comments

No Known Activations