INDEX

Explanations

modal verbs indicating suggestions or recommendations

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

queryInterface

-0.42

 realises

-0.41

Pautan

-0.39

 crochê

-0.37

 realizes

-0.37

 hole

-0.37

Pace

-0.36



-0.36

المكان

-0.36

 findOne

-0.36

POSITIVE LOGITS

Should

0.89

 Should

0.84

should

0.83

 SHOULD

0.76

 should

0.72

 ſhould

0.67

hould

0.65

SHO

0.59

ShouldBe

0.57

 Advice

0.56

Activations Density 0.022%

modal verbs indicating suggestions or recommendations

No Comments

No Known Activations