INDEX

Explanations

phrases related to healthcare, education, and support services

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

 informée

-0.44

 intptr

-0.43

 surla

-0.43

alore

-0.42

IsMutable

-0.40

èdia

-0.40

GenerationType

-0.39

 ويكيميديا

-0.39

SBATCH

-0.38

kegaard

-0.38

POSITIVE LOGITS

titleMargin

0.51

 JSONException

0.42

+#+

0.42

yntaxException

0.40

:+:

0.37

----</

0.37

 fascism

0.37

 declining

0.36

gink

0.36

windowFixed

0.36

Activations Density 0.030%

phrases related to healthcare, education, and support services

No Comments

No Known Activations