INDEX

Explanations

references to coding functions and usage instructions

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

code examples and usage

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

parsedMessage

-0.65

apimachinery

-0.48

=$?

-0.47

RTEX

-0.44

ManagedBean

-0.42

jspb

-0.41

AndEndTag

-0.41

MLLoader

-0.41

يكب

-0.40

ImageContext

-0.40

POSITIVE LOGITS

 Practice

0.47

Practice

0.43

Ceci

0.41

 Usage

0.40

 ddelweddau

0.40

 example

0.40

ApplicationTests

0.40

 usage

0.39

Cecil

0.39

 PRACTICE

0.38

Activations Density 0.556%

references to coding functions and usage instructions

code examples and usage

No Comments

No Known Activations