INDEX

Explanations

indications of deception or misleading information

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

忑

-0.40

省市镇

-0.38

脚注の使い方

-0.38

tagHelperRunner

-0.37

SharedDtor

-0.37

writeFieldEnd

-0.36

latego

-0.32

WebMethod

-0.31

 Monfieur

-0.31

 Schönheit

-0.31

POSITIVE LOGITS

 faked

0.72

 faking

0.71

 phony

0.70

 pretence

0.68

 pretense

0.67

 pretended

0.66

 pretend

0.66

claim

0.64

 fake

0.63

 bogus

0.63

Activations Density 0.480%

indications of deception or misleading information

No Comments

No Known Activations