INDEX

Explanations

references to investigations and discussions involving an ongoing investigation

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

 avoient

-0.65

 parlando

-0.64

 supérieures

-0.62

avaient

-0.58

tuvo

-0.57

 automatiques

-0.57

 enfans

-0.56

 hjär

-0.55

 käns

-0.53

 plastiques

-0.53

POSITIVE LOGITS

 prepares

1.14

 prepare

0.97

 progresses

0.96

 nears

0.92

 continues

0.90

 continue

0.85

 grows

0.79

 progress

0.76

prepare

0.75

 matures

0.73

Activations Density 0.220%

references to investigations and discussions involving an ongoing investigation

No Comments

No Known Activations