INDEX

Explanations

abstract concepts related to conflict or confrontation

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

 constitu

-0.61

 Places

-0.56

Seb

-0.54

 Principles

-0.54

eger

-0.53

ĺħ

-0.53

cale

-0.52

IZE

-0.52

 Pione

-0.50

 Engineers

-0.50

POSITIVE LOGITS

away

1.06

ings

1.00

aways

0.95

offs

0.92

outs

0.91

athon

0.88

back

0.87

up

0.87

backs

0.85

out

0.83

Activations Density 9.171%

abstract concepts related to conflict or confrontation

No Comments

No Known Activations