INDEX

Explanations

strategies and interventions aimed at improving health and social outcomes

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

followed by "to increase"

np_acts-logits-general · gemini-2.0-flash

to increase

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 EconPapers

-0.66

MockMvc

-0.63

EventLoop

-0.60

esez

-0.58

enumi

-0.58

ftagPool

-0.57

تقاوى

-0.57

 esternos

-0.54

Referências

-0.53

 pretending

-0.53

POSITIVE LOGITS

 address

1.08

 enhance

1.04

 promote

1.04

 improve

0.97

 foster

0.90

 Address

0.88

address

0.88

 enhancing

0.87

 boost

0.87

 enable

0.85

Activations Density 0.576%

strategies and interventions aimed at improving health and social outcomes

followed by "to increase"

to increase

No Comments

No Known Activations