INDEX

Explanations

phrases indicating likelihood or expectation

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

 personnalisée

-0.52

RTEE

-0.52

 bParam

-0.51

ArrowToggle

-0.50

 enseña

-0.49

 chrétien

-0.49

stdbool

-0.48

ínsula

-0.47

 říká

-0.47

nodoc

-0.47

POSITIVE LOGITS

be

0.93

 will

0.87

 should

0.64

 likely

0.63

 shall

0.62

 have

0.61

 probably

0.61

 deverá

0.59

]}

0.58

 terão

0.58

Activations Density 0.023%

phrases indicating likelihood or expectation

No Comments

No Known Activations