INDEX

Explanations

terms related to commands and system operations

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

legal and technical terms

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

aly

-0.41

net

-0.39

ver

-0.39

-0.38

 dedicated

-0.38

net

-0.36

 among

-0.36

Ums

-0.36

Ans

-0.35

POSITIVE LOGITS

CloseOperation

0.84

WriteTagHelper

0.83

хьтан

0.80

 Taktlose

0.76

PreInfinity

0.75

 виправивши

0.75

dafx

0.74

InputBorder

0.73

 مشين

0.69

expandindo

0.69

Activations Density 0.138%

terms related to commands and system operations

legal and technical terms

No Comments

No Known Activations