INDEX

Explanations

the word "won" and its variations, indicating a focus on victory or achievement contexts

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

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

gridLayout

-0.56



-0.54

$_['

-0.53

DPR

-0.53

ciaż

-0.53

 EdgeInsets

-0.52

GeoNames

-0.50

ESF

-0.50

ECB

-0.50

tagHelper

-0.50

POSITIVE LOGITS

won

1.14

Won

1.09

won

1.09

Won

1.01

WON

0.95

WON

0.88

 wouldn

0.73

wouldn

0.69

 onCancelled

0.65

 pouvoit

0.64

Activations Density 0.007%

the word "won" and its variations, indicating a focus on victory or achievement contexts

No Comments

No Known Activations