INDEX

Explanations

expressions related to repair and rehabilitation processes

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

fixing or recovering

np_acts-logits-general · gemini-2.0-flash

recovery, repair, and rehabilitation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 24-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_24/width_16k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

AndEndTag

-0.55

SequentialGroup

-0.53

aarrggbb

-0.48

 nahilalakip

-0.47

adag

-0.46

➟

-0.42

complexContent

-0.41

 Kindes

-0.40

Iw

-0.39

orap

-0.39

POSITIVE LOGITS

 recoverable

0.56

 propOrder

0.54

 repair

0.54

なんとか

0.51

挽回

0.50

repair

0.49

 possibility

0.49

ihnachten

0.49

 reinstatement

0.49

何とか

0.48

Activations Density 0.543%

expressions related to repair and rehabilitation processes

fixing or recovering

recovery, repair, and rehabilitation

No Comments

No Known Activations