INDEX

Explanations

language related to threats and violence

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

threatening words

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

InitVars

-0.49

IsMutable

-0.47

tonsoft

-0.46

 विश्वसनीयता

-0.44

∬

-0.44

цездатний

-0.42

Архівовано

-0.40

تقاوى

-0.40

 useAppContext

-0.39

Rptr

-0.39

POSITIVE LOGITS

 threats

0.71

Threats

0.65

 Threats

0.65

threat

0.63

 threatening

0.60

 threat

0.59

 amenazas

0.58

Threat

0.57

 Threat

0.57

 угро

0.57

Activations Density 0.024%

language related to threats and violence

threatening words

No Comments

No Known Activations