INDEX

Explanations

URLs and web-related components

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

Username components or identifiers in chat logs and online communication, particularly in IRC-style formats.

eleuther_acts_top20 · claude-3-7-sonnet-20250219 Triggered by @lu-christina

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Kjelder

-0.66

قایناق‌لار

-0.65

 nakalista

-0.55

 betweenstory

-0.54

AndroidJUnit

-0.53

Enf

-0.50

cof

-0.48

 conten

-0.48

Спољашње

-0.47

Spoljašnje

-0.46

POSITIVE LOGITS

 surla

0.65

 Photocase

0.47

 potrafi

0.42

 referrerpolicy

0.41

LookAnd

0.40

InputBorder

0.38

 hembra

0.36

 NUKAT

0.36

 femininas

0.35

èdia

0.33

Activations Density 0.040%

URLs and web-related components

Username components or identifiers in chat logs and online communication, particularly in IRC-style formats.

No Comments

No Known Activations