INDEX

Explanations

terms related to 'internal' structures and processes

oai_token-act-pair · gpt-4o-mini Triggered by @bot

references to internal aspects of a system or organization, often in technical contexts and frequently contrasted with external elements.

oai_token-act-pair · gpt-5 Triggered by @jyhe0408

The neuron activates strongly for the word "internal" when it appears in formal or technical contexts, particularly when followed by nouns describing components, processes, or structures within systems or organizations.

oai_token-act-pair · claude-4-5-sonnet Triggered by @jyhe0408

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 avoient

-0.68

 queſta

-0.67

 étoient

-0.65

WaitGroup

-0.63

///</

-0.63

BufferException

-0.63

ésult

-0.62

 müſſen

-0.61

 pleaſure

-0.60

étoit

-0.59

POSITIVE LOGITS

 internal

1.30

Internal

1.23

 inside

1.22

 Internal

1.20

internal

1.17

Inside

1.13

inside

1.11

 Inside

1.09

 INTERNAL

1.09

INTERNAL

1.07

Activations Density 0.091%

terms related to 'internal' structures and processes

references to internal aspects of a system or organization, often in technical contexts and frequently contrasted with external elements.

The neuron activates strongly for the word "internal" when it appears in formal or technical contexts, particularly when followed by nouns describing components, processes, or structures within systems or organizations.

No Comments

No Known Activations