INDEX

Explanations

references to bans and prohibitions

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

prohibited or banned

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

ритори

-0.49

nextProps

-0.41

 Fletcher

-0.39

Bou

-0.38

 obra

-0.38

Actors

-0.38

 besoin

-0.38

thought

-0.37

Simply

-0.37

Subviews

-0.37

POSITIVE LOGITS

 banned

0.88

 prohi

0.85

 prohibited

0.82

 banning

0.79

 prohibit

0.77

 prohibiting

0.77

 prohibitions

0.77

 prohibition

0.76

banned

0.75

 forbade

0.75

Activations Density 0.020%

references to bans and prohibitions

prohibited or banned

No Comments

No Known Activations