INDEX

Explanations

various forms of the word "unlock."

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

unlocking and revealing

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

Slf

-0.54

 للمعارف

-0.54

पया

-0.52

 betrekking

-0.46

 StatefulWidget

-0.45

évaluateur

-0.44

tagHelper

-0.44

twimg

-0.42

════

-0.42

ọi

-0.42

POSITIVE LOGITS

 Unlock

0.82

 unlock

0.77

Unlock

0.74

UNLOCK

0.71

 unlocked

0.71

 unpacked

0.69

unlock

0.68

 Unmarshal

0.68

 Unlocked

0.68

 unpack

0.68

Activations Density 0.011%

various forms of the word "unlock."

unlocking and revealing

No Comments

No Known Activations