INDEX

Explanations

common words

np_max-act · gemini-2.0-flash

This neuron fires on evaluative modifiers—particularly adverbs and adjectives (e.g. “definitely,” “certainly,” “most,” “stronger,” “achievement”) that express emphasis or subjective judgment.

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

Associ

-0.07

 annex

-0.07

 incluso

-0.07

Vin

-0.07

 برد

-0.07

 ближ

-0.06

 öğret

-0.06

 nevid

-0.06

<Order

-0.06

我们

-0.06

POSITIVE LOGITS

-в

0.07

+k

0.07

�

0.07

ulumi

0.07

$k

0.06

Kw

0.06

wizard

0.06

 discharged

0.06

akis

0.06

agers

0.06

Activations Density 0.058%

common words

This neuron fires on evaluative modifiers—particularly adverbs and adjectives (e.g. “definitely,” “certainly,” “most,” “stronger,” “achievement”) that express emphasis or subjective judgment.

No Comments

No Known Activations

common words

This neuron fires on evaluative modifiers—particularly adverbs and adjectives (e.g. “definitely,” “certainly,” “most,” “stronger,” “achievement”) that express emphasis or subjective judgment.

No Comments

No Known Activations