INDEX

Explanations

terms related to crime and misconduct

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

crimes and unlawful acts

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

Попис

-0.44

 aanges

-0.40

 tärke

-0.36

 Kerja

-0.36

 gekomen

-0.36

 fédé

-0.35

wsdl

-0.35

 kese

-0.34

 trekken

-0.34

 clín

-0.34

POSITIVE LOGITS

 MainAxisSize

0.63

AndEndTag

0.60

IsMutable

0.57

Theft

0.55

 ***!

0.55

terrorism

0.53

 Taktlose

0.53

 Theft

0.53

MigrationBuilder

0.52

theft

0.52

Activations Density 0.050%

terms related to crime and misconduct

crimes and unlawful acts

No Comments

No Known Activations