INDEX

Explanations

terms related to intervention programs

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

 Keuangan

-0.45

untungan

-0.44

Sounds

-0.42

 Umum

-0.41

 Inscrivez

-0.40

 Suivez

-0.40

 défauts

-0.40

 zijne

-0.40

 Grecs

-0.40

dür

-0.39

POSITIVE LOGITS

 intervention

1.15

 interventions

1.14

 Intervention

1.00

intervention

1.00

 intervene

1.00

Intervention

0.97

 Interventions

0.95

 interven

0.92

Interventions

0.84

 intervened

0.82

Activations Density 0.017%

terms related to intervention programs

No Comments

No Known Activations