INDEX

Explanations

phrases indicating legal actions and outcomes

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

discontinuation or ending

np_acts-logits-general · gemini-2.0-flash

endings and failures

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

ArgsConstructor

-0.85

 Réponses

-0.85

 estekak

-0.80

 שוליים

-0.77

rungsseite

-0.73

сылкі

-0.71

hoeddwyd

-0.69

anyahu

-0.68

SBATCH

-0.66

 poveznice

-0.65

POSITIVE LOGITS

掉

0.74

 abruptly

0.72

 soon

0.66

 altogether

0.60

 uncer

0.60

 premat

0.59

 faute

0.59

 gracefully

0.59

 completely

0.58

 after

0.57

Activations Density 0.546%

phrases indicating legal actions and outcomes

discontinuation or ending

endings and failures

No Comments

No Known Activations