INDEX

Explanations

phrases indicating causation and influence within complex contexts

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 36-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_36/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.36.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 ModelExpression

-0.49

олові

-0.45

 noDo

-0.44

UrlResolution

-0.44

nextProps

-0.43

Cyfarwyddwr

-0.42

Collegamenti

-0.41

webElementXpaths

-0.41

DIPSETTING

-0.41

Externé

-0.40

POSITIVE LOGITS

 partly

2.58

 partially

2.31

partly

2.25

 Partly

2.19

Partly

2.05

 Partially

1.97

 parcialmente

1.63

 partial

1.59

 largely

1.46

parti

1.43

Activations Density 0.598%

phrases indicating causation and influence within complex contexts

No Comments

No Known Activations