INDEX

Explanations

references to knives and stabbing actions

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

knife, knives

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

 Greet

-0.53

sqcup

-0.49

 Beal

-0.49

 Urqu

-0.48

 Popp

-0.48

Eas

-0.47

 Leighton

-0.47

 Beatty

-0.46

 Prosper

-0.46

Erb

-0.46

POSITIVE LOGITS

 Knife

1.33

 knife

1.32

Knife

1.28

knife

1.26

 knives

1.13

 Knives

1.11

 couteau

1.03

 cuchillo

0.92

Kni

0.91

kni

0.73

Activations Density 0.004%

references to knives and stabbing actions

knife, knives

No Comments

No Known Activations