INDEX

Explanations

references to distinct concepts, items, or categories that are treated individually or separately

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

Separation, independence, or isolation

np_acts-logits-general · gemini-2.0-flash

stand alone, separately, or independently

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

WriteBarrier

-0.58

/*

-0.55

 >=",

-0.54

PerformLayout

-0.53

AndEndTag

-0.52

AddHtmlAttribute

-0.50

formaz

-0.49

ListItemIcon

-0.48

 pinulongan

-0.48

POSITIVE LOGITS

 separate

2.33

 independent

2.32

 separately

2.14

 Separate

2.13

 standalone

2.11

 independently

2.10

separate

2.10

独立

2.09

Separate

2.08

 seperate

2.03

Activations Density 0.773%

references to distinct concepts, items, or categories that are treated individually or separately

Separation, independence, or isolation

stand alone, separately, or independently

No Comments

No Known Activations