INDEX

Explanations

terms related to medical contexts, particularly in relation to healthcare scenarios and procedures

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_33/width_131k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.33.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

awtextra

-0.65

expandindo

-0.55

 noDo

-0.52

 betweenstory

-0.47

sidemargin

-0.46

Personensuche

-0.45

enderror

-0.45

AutoresizingMask

-0.43

WriteTagHelper

-0.42

dibles

-0.41

POSITIVE LOGITS

 elsewhere

0.67

 also

0.65

 även

0.63

 specifically

0.63

 myös

0.60

 sogar

0.59

 even

0.58

 separately

0.58

เฉพาะ

0.55

 indirectly

0.55

Activations Density 0.367%

terms related to medical contexts, particularly in relation to healthcare scenarios and procedures

No Comments

No Known Activations