INDEX

Explanations

expressions of negation or denial in experiences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 38-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_38/width_16k/average_l0_128

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.38.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

 rồi

-0.46

 also

-0.45

already

-0.44

 Already

-0.43

also

-0.42

 already

-0.40

Already

-0.40

definitely

-0.40

 también

-0.40

 també

-0.39

POSITIVE LOGITS

 bothered

1.11

 bothering

1.10

 bother

1.08

 ever

1.07

 even

0.96

 siquiera

0.94

 bothers

0.94

 EVER

0.88

 budge

0.86

 cared

0.77

Activations Density 0.683%

expressions of negation or denial in experiences

No Comments

No Known Activations