INDEX

Explanations

markup and formatting elements

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

in-text citations with abbreviations/genes

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

IntoConstraints

-0.72

 betweenstory

-0.70

 MainAxisSize

-0.69

Välislingid

-0.62

WebElementEntity

-0.61

 queſta

-0.59

 DBNull

-0.59

DockStyle

-0.59

tagext

-0.57

 startIndex

-0.57

POSITIVE LOGITS

<em>

1.65

<i>

1.63

<u>

1.08

<strong>

1.05

<b>

1.02

__(/*!

0.92

<code>

0.81

<sub>

0.73

/*

0.70

“

0.66

Activations Density 0.022%

markup and formatting elements

in-text citations with abbreviations/genes

No Comments

No Known Activations