INDEX

Explanations

az

np_max-act · gemini-2.0-flash

This neuron detects structural markup and metadata tokens (e.g. speaker tags and delimiters) rather than content 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

ρός

-0.07

리를

-0.07

做

-0.07

 hike

-0.07

ременно

-0.07

Fin

-0.06

thickness

-0.06

Porno

-0.06

Dean

-0.06

 Impl

-0.06

POSITIVE LOGITS

OPTION

0.07

김

0.07

<option

0.06

ocr

0.06

activate

0.06

 resort

0.06

.kafka

0.06

 substantive

0.06

ắ

0.06

mma

0.06

Activations Density 0.005%

az

This neuron detects structural markup and metadata tokens (e.g. speaker tags and delimiters) rather than content words.

No Comments

No Known Activations