INDEX

Explanations

phrases discussing efficacy and assessments of policies or treatments

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

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

 either

-0.80

Either

-0.79

 even

-0.78

either

-0.77

 Either

-0.77

even

-0.71

 prostu

-0.65

 inoltre

-0.65

 invece

-0.65

 entweder

-0.65

POSITIVE LOGITS

 technically

1.31

 nominally

1.17

 ostensibly

1.08

 outwardly

1.03

 admittedly

1.00

 theoretically

0.99

 superfic

0.96

 initially

0.92

may

0.91

 téc

0.88

Activations Density 0.504%

phrases discussing efficacy and assessments of policies or treatments

No Comments

No Known Activations