INDEX

Explanations

legal and procedural terminology related to convictions and assignments of error

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 pinulongan

-0.72

 مشين

-0.65

WithIOException

-0.57

Rüyada

-0.55

ConstraintMaker

-0.53

:][

-0.50

ðsíða

-0.50

ſicht

-0.50

󠁴

-0.49

الدراسه

-0.49

POSITIVE LOGITS

Another

0.44

another

0.42

 Another

0.40

richTextPanel

0.36

これも

0.36

 weiterer

0.34

 another

0.33

Again

0.32

again

0.32

 tích

0.32

Activations Density 0.854%

legal and procedural terminology related to convictions and assignments of error

No Comments

No Known Activations