INDEX

Explanations

phrases associated with critiques of actions, especially regarding negligence and performance in various 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

parsedMessage

-0.68

 zvuky

-0.60

rungsseite

-0.55

makeText

-0.55

Obrázky

-0.50

 تعدى

-0.48

 '\\;'

-0.48

➯

-0.48

 onCreateView

-0.48

 Esau

-0.48

POSITIVE LOGITS

 mismanagement

0.47

 inaction

0.44

 failing

0.42

 failings

0.40

 failure

0.39

 défa

0.37

 delay

0.37

 mantenga

0.37

 lapses

0.36

 delays

0.35

Activations Density 0.049%

phrases associated with critiques of actions, especially regarding negligence and performance in various contexts

No Comments

No Known Activations