INDEX

Explanations

formal legal and policy terminology

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

specifies policy or guidelines

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

 '\\;'

-0.82

GEBURTSDATUM

-0.75

цездатний

-0.73

ValueStyle

-0.71

 EconPapers

-0.70

 CreateTagHelper

-0.70

 OMITBAD

-0.70

MigrationBuilder

-0.70

 Савезне

-0.69

 linkovi

-0.69

POSITIVE LOGITS

 provide

0.38

 guide

0.36

 provides

0.35

 information

0.34

 serves

0.33

 covers

0.29

 intended

0.29

 intend

0.29

ここでは

0.28

 serve

0.28

Activations Density 0.019%

formal legal and policy terminology

specifies policy or guidelines

No Comments

No Known Activations