INDEX

Explanations

court decisions

np_max-act · gemini-2.0-flash

This neuron detects boilerplate phrases indicating a summary affirmance by an appellate authority (e.g. “summarily affirmed”).

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

exact

-0.08

AT

-0.07

Sources

-0.07

 moderator

-0.07

 slogans

-0.06

_LAT

-0.06

OBJ

-0.06

Ara

-0.06

ctx

-0.06

架

-0.06

POSITIVE LOGITS

cripts

0.07

 flooring

0.06

clo

0.06

 nedeni

0.06

.AutoScaleMode

0.06

dın

0.06

 knockout

0.06

stery

0.06

 електрон

0.06

_import

0.06

Activations Density 0.013%

court decisions

This neuron detects boilerplate phrases indicating a summary affirmance by an appellate authority (e.g. “summarily affirmed”).

No Comments

No Known Activations