INDEX

Explanations

references to decapitation and severed heads

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

head movements and states

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

 strøm

-0.45

 архивлан

-0.45

SpringRunner

-0.42

 Pact

-0.41

 Италијани

-0.40

 iprot

-0.40

mbggenerated

-0.40

 varandra

-0.39

 føl

-0.39

GroupLayout

-0.38

POSITIVE LOGITS

 head

1.00

 heads

0.95

 Kopf

0.91

 HEAD

0.89

Head

0.87

 Head

0.84

head

0.82

 Heads

0.82

 header

0.79

 cabeza

0.79

Activations Density 0.055%

references to decapitation and severed heads

head movements and states

No Comments

No Known Activations