INDEX

Explanations

specific dates or numerical sequences expressed in number-word formats

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

âĿ

-1.08

tions

-0.93

 Cruiser

-0.93

letter

-0.90

 McCann

-0.89

Ont

-0.87

HP

-0.85

mails

-0.83

©¶æ¥µ

-0.83

sheets

-0.83

POSITIVE LOGITS

 ascertain

1.09

ggles

1.09

wing

1.06

 prevent

1.04

 minimize

1.03

yx

1.02

 date

1.00

wered

0.99

 ensure

0.98

 remedy

0.98

Activations Density 0.865%

specific dates or numerical sequences expressed in number-word formats

No Comments

No Known Activations