INDEX

Explanations

terms related to medical imaging and pathology analysis

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

 betweenstory

-0.61

 protoimpl

-0.57

 faſt

-0.56

Personendaten

-0.54

Carriera

-0.52

⟬

-0.50

inSlope

-0.50

 صوتيه

-0.49

ſelves

-0.48

Xna

-0.47

POSITIVE LOGITS

 determination

2.97

 Determination

2.63

 determine

2.58

 determining

2.52

 determinations

2.48

determination

2.48

Determination

2.42

determin

2.41

 determined

2.39

determine

2.36

Activations Density 0.043%

terms related to medical imaging and pathology analysis

No Comments

No Known Activations