INDEX

Explanations

to be

np_max-act · gemini-2.0-flash

The neuron lights up on content-heavy, domain-specific nouns and technical terms (e.g. “cells,” “methods,” “analysis,” “procedures”) that often appear in scientific writing.

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

trained

-0.07

SCO

-0.07

anglicky

-0.06

цю

-0.06

 monitored

-0.06

version

-0.06

 gamer

-0.06

 Sharing

-0.06

 Mustang

-0.06

demo

-0.06

POSITIVE LOGITS

is

0.08

—is

0.08

Was

0.08

was

0.07

-small

0.07

 MANUAL

0.07

are

0.06

|RF

0.06

Is

0.06

-is

0.06

Activations Density 0.208%

to be

The neuron lights up on content-heavy, domain-specific nouns and technical terms (e.g. “cells,” “methods,” “analysis,” “procedures”) that often appear in scientific writing.

No Comments

No Known Activations

to be

The neuron lights up on content-heavy, domain-specific nouns and technical terms (e.g. “cells,” “methods,” “analysis,” “procedures”) that often appear in scientific writing.

No Comments

No Known Activations