INDEX

Explanations

mentions of executive roles or titles

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

ợt

-0.42

 adesivos

-0.38

awtextra

-0.38

 nakalista

-0.38

 wystaw

-0.37

 montaje

-0.35

/***/

-0.35

 equipamiento

-0.35

밌

-0.34

 všem

-0.34

POSITIVE LOGITS

 summary

0.58

 summaries

0.55

 Summary

0.55

summary

0.52

 suites

0.48

 SUMMARY

0.48

 Exec

0.47

 suite

0.47

 committee

0.46

 producer

0.46

Activations Density 0.134%

mentions of executive roles or titles

No Comments

No Known Activations