INDEX

Explanations

phrases or sentences related to putting information on record or making official statements

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Configuration

neuronpedia/gpt2-small__res_scefr-ajt/6-res_scefr-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

ÅĤ

-1.10

erva

-1.10

hov

-1.05

endar

-1.05

emouth

-1.04

arcity

-1.02

igue

-0.96

wake

-0.95

monton

-0.95

nesota

-0.94

POSITIVE LOGITS

 confess

1.04

 regarding

0.97

 negotiations

0.95

 ABOUT

0.95

 discussions

0.93

 beforehand

0.92

notes

0.92

 pledging

0.91

 discussing

0.91

 announcing

0.91

Activations Density 0.779%

phrases or sentences related to putting information on record or making official statements

No Comments

No Known Activations