INDEX

Explanations

phrases indicating validation or testing of concepts and data

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 36-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_36/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.36.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

jména

-0.47

 Betrachtung

-0.46

idać

-0.45

PyExc

-0.45

슷

-0.44

 cérémonie

-0.42

 réglementation

-0.40

occasione

-0.40

 discussions

-0.40

 InputDecoration

-0.40

POSITIVE LOGITS

 validity

1.84

 correctness

1.51

validity

1.45

 legitimacy

1.42

 viability

1.42

 Validity

1.40

 veracity

1.39

 adequacy

1.38

 authenticity

1.35

 suitability

1.30

Activations Density 1.034%

phrases indicating validation or testing of concepts and data

No Comments

No Known Activations