INDEX

Explanations

terms related to primary care and medical providers

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

medical professionals

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 disparu

-0.40

lappe

-0.40

Cra

-0.37

 arşivlendi

-0.35

 disfr

-0.35

 escaped

-0.34

 suspendu

-0.33

 lendemain

-0.32

anniversaire

-0.32

 bonheur

-0.32

POSITIVE LOGITS

 physicians

0.78

 physician

0.71

 clinicians

0.66

 وتسجيلات

0.64

 Physician

0.64

 prescribing

0.63

 prescri

0.60

Physician

0.60

 Physicians

0.60

sizeCache

0.60

Activations Density 0.087%

terms related to primary care and medical providers

medical professionals

No Comments

No Known Activations