INDEX

Explanations

descriptive actions and interactions among different parties within a story or news event

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Configuration

neuronpedia/gpt2-small__res_scl-ajt/6-res_scl-ajt

Prompts (Dashboard)

12,288 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

46,080

Data Type

torch.float32

Hook Point

blocks.6.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

apollo-research/Skylion007-openwebtext-tokenizer-gpt2

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Mages

-0.59

 unemploy

-0.57

 Sorceress

-0.57

Oracle

-0.54

.",

-0.52

 contributor

-0.52

 retirees

-0.51

 attackers

-0.51

 Rolls

-0.51

 nailed

-0.50

POSITIVE LOGITS

appropriately

0.95

 earnest

0.87

 vain

0.80

clus

0.80

humane

0.78

versely

0.76

clusively

0.74

odes

0.72

aus

0.71

orously

0.69

Activations Density 19.581%

descriptive actions and interactions among different parties within a story or news event

No Comments

No Known Activations