INDEX

Explanations

expressions related to threats and intimidation

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

say or do certain things

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

 voulu

-0.42

 especializados

-0.41

zieży

-0.41

 egli

-0.41

 especializado

-0.39

ajur

-0.39

 inconnu

-0.39

șit

-0.39

ähteet

-0.38

worpen

-0.38

POSITIVE LOGITS

 ujednoznacz

0.54

Portály

0.47

 Hashtable

0.47

windowFixed

0.45

 ModelExpression

0.43

+#+#

0.43

insee

0.43

 betweenstory

0.42

KommentareTeilen

0.41

Activations Density 0.064%

expressions related to threats and intimidation

say or do certain things

No Comments

No Known Activations