INDEX

Explanations

references to dangers and risks posed by various entities or circumstances

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

danger, threat, risk

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

 wireType

-0.45

 memoized

-0.44

 sesiones

-0.44

껏

-0.43

lasty

-0.40

onaut

-0.40

OGND

-0.40

IVATE

-0.40

 stuks

-0.40

GOTREF

-0.39

POSITIVE LOGITS

 threat

1.21

 risks

1.13

 danger

1.12

 risk

1.12

 threats

1.11

 dangers

1.11

threat

1.05

 Risks

1.02

 Threat

1.02

Threat

1.00

Activations Density 0.041%

references to dangers and risks posed by various entities or circumstances

danger, threat, risk

No Comments

No Known Activations