INDEX

Explanations

references to live performances or recordings

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

inSlope

-0.57

 Autorizaciones

-0.52

 ujednoznacz

-0.48

IntoConstraints

-0.46

 Autorisations

-0.46

diseño

-0.46

Personensuche

-0.45

Aiheesta

-0.44

 initState

-0.44

 createSlice

-0.44

POSITIVE LOGITS

 recordings

0.54

Recorded

0.51

 recorded

0.49

 Recorded

0.49

 recording

0.47

recorded

0.44

 live

0.42

 filmed

0.41

 запи

0.40

Recording

0.39

Activations Density 0.009%

references to live performances or recordings

No Comments

No Known Activations