INDEX

Explanations

terms indicating the initial state or configuration within programming contexts

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

 للمعارف

-0.81

 BrowserModule

-0.69

erçe

-0.68

 bVar

-0.66

 onError

-0.66

 propOrder

-0.66

 müſſen

-0.65

 laſſen

-0.64

ImageContext

-0.64

 getProduct

-0.63

POSITIVE LOGITS

INITIAL

0.79

setInitial

0.76

initial

0.74

Initial

0.73

 Initial

0.61

 initial

0.57

 INITIAL

0.55

Intial

0.52

 нача

0.47

itial

0.47

Activations Density 0.006%

terms indicating the initial state or configuration within programming contexts

No Comments

No Known Activations