INDEX

Explanations

technical terms and identifiers related to web and email protocols

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

Content-Type and headers

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

 murni

-0.58

จบ

-0.47

 superare

-0.45

 crossorigin

-0.44

 pelan

-0.44

 vastaan

-0.44

 gezeigt

-0.44

impresa

-0.43

 decidir

-0.43

Résultats

-0.43

POSITIVE LOGITS

 defaultstate

0.60

">//

0.59

AutoScaleMode

0.58

хьтан

0.52

WriteTagHelper

0.52

 ComVisible

0.50

 transfieras

0.49

 TokenType

0.48

Diwedd

0.47

 autorytatywna

0.46

Activations Density 0.001%

technical terms and identifiers related to web and email protocols

Content-Type and headers

No Comments

No Known Activations