INDEX

Explanations

proper nouns, particularly names of individuals

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

unda

-0.63

ge

-0.62

atively

-0.60

atical

-0.57

opsis

-0.57

gments

-0.55

ctors

-0.53

ctor

-0.53

arily

-0.53

gebra

-0.53

POSITIVE LOGITS

hips

0.79

hip

0.75

hops

0.72

0.65

pring

0.62

mith

0.62

peed

0.60

hire

0.59

boro

0.59

heet

0.59

Activations Density 7.533%

proper nouns, particularly names of individuals

No Comments

No Known Activations