INDEX

Explanations

words related to news articles and technical details

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

ãĥ¼ãĥĨãĤ£

-0.67

flies

-0.64

acity

-0.63

uine

-0.61

keepers

-0.60

ocide

-0.59

riott

-0.58

hist

-0.58

atur

-0.58

alist

-0.56

POSITIVE LOGITS

FTWARE

0.89

 READ

0.87

LINE

0.86

 COVER

0.77

GROUND

0.75

ILE

0.74

ES

0.74

LOAD

0.73

LIST

0.73

ING

0.72

Activations Density 7.701%

words related to news articles and technical details

No Comments

No Known Activations