INDEX

Explanations

phrases related to deception and manipulation, particularly in the context of politics and personal relationships

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 10-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_10/width_16k/average_l0_77

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.10.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

DotNetBar

-0.64

 newBuilder

-0.63

المناصب

-0.55

 Wiktionnaire

-0.55

 unification

-0.55

 pleaſure

-0.54

RunAsync

-0.53

 Universitaria

-0.52

 effusion

-0.52

 crossorigin

-0.52

POSITIVE LOGITS

 pretend

0.74

 pretended

0.64

 fake

0.63

 pretending

0.61

 pretends

0.59

 faked

0.56

tagHelper

0.53

 giả

0.52

虚

0.52

fake

0.51

Activations Density 0.403%

phrases related to deception and manipulation, particularly in the context of politics and personal relationships

No Comments

No Known Activations