INDEX

Explanations

patterns of equality and comparison in numeric or logical contexts

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

"equal" or its variants

np_acts-logits-general · gemini-2.0-flash

equality and comparison

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

FontStyle

-0.45

 domov

-0.40

antren

-0.39

HostException

-0.39

Hdr

-0.39

ArgumentParser

-0.39

 Hols

-0.39

-0.38

 turi

-0.38

 researches

-0.38

POSITIVE LOGITS

Equal

1.12

equal

1.09

 equal

1.06

 EQUAL

1.05

 Equal

1.03

EQUAL

1.00

Equals

0.94

 Equality

0.88

 Equals

0.87

equals

0.86

Activations Density 0.229%

patterns of equality and comparison in numeric or logical contexts

"equal" or its variants

equality and comparison

No Comments

No Known Activations