INDEX

Explanations

questions about how to perform tasks or actions

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

Follows "how do I" or "how to"

np_acts-logits-general · gemini-2.0-flash

how to do things

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 asupra

-0.60

aarrggbb

-0.60

Jeografia

-0.59

ViewFeatures

-0.56

новниш

-0.56

principalColumn

-0.55

 insuffisamment

-0.55

ริง

-0.52

 étudi

-0.52

 corrida

-0.51

POSITIVE LOGITS

 best

0.88

 approach

0.80

 Approach

0.74

 safely

0.74

Approach

0.73

 avoid

0.72

 achieve

0.70

approach

0.70

才能

0.69

 overcome

0.69

Activations Density 0.250%

questions about how to perform tasks or actions

Follows "how do I" or "how to"

how to do things

No Comments

No Known Activations