INDEX

Explanations

legal terminology and phrases within formal texts

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

numbered lists and parentheticals

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

RegressionTest

-0.59

 &___

-0.55

:+:

-0.54

istoitu

-0.50

.*")]

-0.47

 BoxFit

-0.46

jspb

-0.46

#+#

-0.46

 invokingState

-0.44

 sorpresas

-0.44

POSITIVE LOGITS

 okuyayım

0.43

הערות

0.43

beginTransaction

0.39

 للاسماء

0.39

VersionUID

0.38

Koa

0.36

msgTypes

0.34

عرِّف

0.34

 Baron

0.34

 ſch

0.34

Activations Density 0.067%

legal terminology and phrases within formal texts

numbered lists and parentheticals

No Comments

No Known Activations