INDEX

Explanations

the term "prior" and its variants, indicating a focus on previous events or conditions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

 restTemplate

-0.64

 ProtoMessage

-0.62

GOTREF

-0.61

""}

-0.59

ciaż

-0.59

|/

-0.57

RUnlock

-0.56

ędzynarod

-0.55

ThroughAttribute

-0.54

epam

-0.53

POSITIVE LOGITS

PRIOR

0.73

prior

0.69

front

0.65

itisation

0.65

vori

0.65

 PRIOR

0.63

 preg

0.62

 Mitar

0.62

before

0.61

itize

0.61

Activations Density 0.093%

the term "prior" and its variants, indicating a focus on previous events or conditions

No Comments

No Known Activations