INDEX

Explanations

narratives involving conflict and power dynamics among characters

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

destruction or overthrowing

np_acts-logits-general · gemini-2.0-flash

overthrow or destroy

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

usehen

-0.53

aleg

-0.49

TagHelper

-0.47

matchCondition

-0.46

 diken

-0.45

новништво

-0.44

 consenti

-0.44

AnchorTagHelper

-0.43

ModelSerializer

-0.43

Answered

-0.42

POSITIVE LOGITS

 destroy

1.12

 destroying

1.05

 kill

1.00

 overthrow

0.95

 destroys

0.95

 obliter

0.91

 killing

0.88

 oust

0.86

 defeating

0.86

 annihil

0.85

Activations Density 0.662%

narratives involving conflict and power dynamics among characters

destruction or overthrowing

overthrow or destroy

No Comments

No Known Activations