INDEX

Explanations

references to explosive devices or bombs

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

bombs and grenades

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

lewati

-0.54

 umane

-0.52

 diferenças

-0.50

 oczeki

-0.50

 murni

-0.48

 față

-0.48

ViewInit

-0.48

 umani

-0.47

 Fairview

-0.47

 diferença

-0.47

POSITIVE LOGITS

 bomb

1.30

 Bomb

1.25

bomb

1.17

Bomb

1.16

 bombs

1.13

 BOMB

1.11

 Bombs

1.05

bombs

1.03

 bombing

1.02

 bombe

0.95

Activations Density 0.004%

references to explosive devices or bombs

bombs and grenades

No Comments

No Known Activations