INDEX

Explanations

references to time frames or sequences in events

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

Preceding tokens of "follows" or "followed"

np_acts-logits-general · gemini-2.0-flash

follows appeal

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

KommentareTeilen

-0.49

__(/*!

-0.49

Pranala

-0.47

CreateMap

-0.45

órmula

-0.44

一面

-0.42

yargs

-0.41

 druge

-0.41

منابع

-0.41

anone

-0.41

POSITIVE LOGITS

 follow

2.63

 follows

2.45

 followed

2.41

 FOLLOW

2.32

follow

2.27

Follow

2.23

 Follow

2.20

followed

2.04

FOLLOW

2.02

 Followed

1.99

Activations Density 0.385%

references to time frames or sequences in events

Preceding tokens of "follows" or "followed"

follows appeal

No Comments

No Known Activations