INDEX

Explanations

phrases and words related to agreement and consensus

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

stating agreement

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_24/width_131k/average_l0_97

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.24.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

 مرئيه

-0.38

putchar

-0.36

 Reynolds

-0.34

الدراسه

-0.33

electron

-0.33

 vide

-0.33

NewUrlParser

-0.33

 pharma

-0.32

 listOf

-0.31

 isActive

-0.31

POSITIVE LOGITS

 agreed

0.86

 Agreed

0.82

Agreed

0.78

Agree

0.78

 agreement

0.76

 agree

0.76

agreed

0.76

 agreeing

0.75

 Agree

0.74

 agrees

0.69

Activations Density 0.077%

phrases and words related to agreement and consensus

stating agreement

No Comments

No Known Activations