INDEX

Explanations

directives related to avoidance and recommendations against certain actions

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

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

NameInMap

-0.68

ьаж

-0.54

 ſind

-0.53

ViewInit

-0.49

BibitemShut

-0.49

=$?

-0.48

matchCondition

-0.48

parsedMessage

-0.47

 Complement

-0.47

 שוליים

-0.46

POSITIVE LOGITS

 Avoid

0.56

Avoid

0.54

 avoid

0.53

avoid

0.52

 avoided

0.50

 AVOID

0.44

 avoiding

0.44

 interdit

0.43

 warned

0.42

heim

0.41

Activations Density 0.046%

directives related to avoidance and recommendations against certain actions

No Comments

No Known Activations