INDEX

Explanations

proper nouns and names of individuals in news articles

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

ulhu

-0.44

ICLE

-0.43

»Ĵ

-0.42

HAEL

-0.42

lished

-0.41

lehem

-0.41

escription

-0.41

ģĸ

-0.41

 corrid

-0.41

ASED

-0.40

POSITIVE LOGITS

horn

0.46

 Dragonbound

0.46

zai

0.42

hirt

0.42

iland

0.41

oxide

0.41

velt

0.40

acht

0.40

 Revenge

0.40

ppa

0.40

Activations Density 11.391%

proper nouns and names of individuals in news articles

No Comments

No Known Activations