INDEX

Explanations

terms related to policies and legal disclosures

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

policies and documents

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

 CURIAM

-0.49

SequentialGroup

-0.48

ScopeManager

-0.47

 EconPapers

-0.47

AddTagHelper

-0.46

AndroidJUnit

-0.46

findpost

-0.45

 marché

-0.45

 propOrder

-0.44

 corduroy

-0.44

POSITIVE LOGITS

contentLoaded

0.45

 rules

0.43

asteroide

0.42

 online

0.41

 marbre

0.40

 printed

0.40

 regulations

0.40

 documents

0.39

 downloaded

0.38

 Winaray

0.38

Activations Density 0.036%

terms related to policies and legal disclosures

policies and documents

No Comments

No Known Activations