INDEX

Explanations

terms that describe low levels or low values across various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 11-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_11/width_16k/average_l0_80

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.11.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

 religieuses

-0.83

 virtù

-0.82

 graphiques

-0.80

 senere

-0.80

 européennes

-0.79

 tekem

-0.78

 relatifs

-0.77

 véhic

-0.77

 supérieurs

-0.76

 suivants

-0.76

POSITIVE LOGITS

LOW

1.68

Low

1.66

low

1.60

Low

1.50

 Lowry

1.41

low

1.31

LOW

1.22

 Lowe

1.16

低

1.14

 lows

1.09

Activations Density 0.094%

terms that describe low levels or low values across various contexts

No Comments

No Known Activations