INDEX

Explanations

the term "over" and its various forms

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

over and under prefixes

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

tagHelperRunner

-0.77

IntoConstraints

-0.72

 AssemblyProduct

-0.71

Gambas

-0.66

principalTable

-0.65

➯

-0.64

fjspx

-0.64

 semelh

-0.64

 Wikimedijinoj

-0.63

AndEndTag

-0.63

POSITIVE LOGITS

0.54

0.48

0.40

，

0.40

getMenuInflater

0.38

’

0.37

fl

0.37

0.36

 قه

0.35

 pié

0.35

Activations Density 0.033%

the term "over" and its various forms

over and under prefixes

No Comments

No Known Activations