INDEX

Explanations

terms and phrases related to healthcare, research designs, and data analysis

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 alvor

-0.35

 oxygen

-0.33

uș

-0.32

 synes

-0.32

 avanzado

-0.32

 oameni

-0.31

pă

-0.31

 forhold

-0.31

 default

-0.31

EnableWeb

-0.30

POSITIVE LOGITS

⟬

0.65

 Paglinawan

0.63

IntoConstraints

0.62

utilisons

0.60

+#+

0.60

ConstraintMaker

0.59

tinyos

0.59

évaluateur

0.58

RTEE

0.58

 FetchType

0.55

Activations Density 0.731%

terms and phrases related to healthcare, research designs, and data analysis

No Comments

No Known Activations