INDEX

Explanations

references to evidence of strength or affirmation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 7-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_7/width_16k/average_l0_69

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.7.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

 >=",

-0.57

 betweenstory

-0.55

ymal

-0.54

rond

-0.52

lji

-0.52

PageContext

-0.50

 Eura

-0.50

arcs

-0.48

ULLY

-0.47

lings

-0.47

POSITIVE LOGITS

 testament

0.81

 témo

0.75

 testifies

0.70

それだけ

0.68

RegressionTest

0.66

 preuve

0.66

 indication

0.66

testify

0.66

 menunjukkan

0.63

 validates

0.62

Activations Density 0.322%

references to evidence of strength or affirmation

No Comments

No Known Activations