INDEX

Explanations

programming-related constructs and object initialization patterns

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

 MethodInfo

-0.38

 flocks

-0.37

 ویکی‌آمباردا

-0.35

AILY

-0.34

coroutines

-0.34

הערות

-0.34

ゅう

-0.33

Complexity

-0.33

 gynhyrchwyd

-0.32

GeneratedCode

-0.32

POSITIVE LOGITS

 informée

0.48

InstrumentedTest

0.47

dė

0.47

BASEPATH

0.46

0.43

IBLES

0.41

indest

0.40

 angew

0.40

Manbalar

0.39

 desmotivaciones

0.39

Activations Density 0.050%

programming-related constructs and object initialization patterns

No Comments

No Known Activations