INDEX

Explanations

words associated with medical treatments, particularly for serious conditions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-mlp-65k

Configuration

google/gemma-scope-2b-pt-mlp/layer_0/width_65k/average_l0_72

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.0.hook_mlp_out

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

himself

-0.60

 himself

-0.59

참고

-0.57

multer

-0.56

aragus

-0.54

róleo

-0.54

 SOCK

-0.54

 Krzyż

-0.54

AUTOR

-0.53

 Morn

-0.52

POSITIVE LOGITS

 nahilalakip

0.74

AddTagHelper

0.69

 NSCoder

0.66

 fewest

0.57

IntoConstraints

0.55

versial

0.53

Попис

0.52

 sportowe

0.52

 Ecclesiastical

0.51

arily

0.51

Activations Density 0.038%

words associated with medical treatments, particularly for serious conditions

No Comments

No Known Activations