INDEX

Explanations

words indicating frequency or consistency in actions and reporting

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

 Houſe

-0.54

 houſe

-0.54

 ſtate

-0.49

 avoient

-0.47

 ſche

-0.47

 createState

-0.47

 enfans

-0.47

 pleaſure

-0.45

 nästa

-0.45

 quæ

-0.45

POSITIVE LOGITS

 continually

0.81

 repeatedly

0.80

 constantly

0.78

 persistently

0.76

 continuously

0.76

marily

0.74

 consistently

0.72

 routinely

0.71

 frequently

0.70

 regularly

0.68

Activations Density 0.013%

words indicating frequency or consistency in actions and reporting

No Comments

No Known Activations