INDEX

Explanations

terms related to "limited" and "limit."

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

ThroughAttribute

-0.49

 Blumen

-0.42

しかった

-0.39

ImageIO

-0.39

 kela

-0.39

 Zwie

-0.39

olera

-0.39

 برانيه

-0.38

 Image

-0.38

 noqa

-0.38

POSITIVE LOGITS

 edition

0.57

 Liability

0.56

 scope

0.54

 liability

0.52

edition

0.51

Liability

0.51

 Edition

0.50

EDEFAULT

0.47

lessly

0.47

 EDITION

0.47

Activations Density 0.267%

terms related to "limited" and "limit."

No Comments

No Known Activations