INDEX

Explanations

instances of negation or contradiction

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

not literally true

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

 nadzieję

-0.49

HasAnnotation

-0.49

bootstrapcdn

-0.48

RenderAtEndOf

-0.46

 szüks

-0.44

orologio

-0.44

 дописавши

-0.43

tagHelperRunner

-0.42

üyada

-0.41

 intptr

-0.40

POSITIVE LOGITS

jooq

0.45

Diweddarwch

0.38

BCG

0.38

컷

0.38

InputTagHelper

0.37

TAC

0.36

 arşivlendi

0.36

 egentligen

0.35

dec

0.35

 XCTest

0.35

Activations Density 0.061%

instances of negation or contradiction

not literally true

No Comments

No Known Activations