INDEX

Explanations

references to legal or judicial terms

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

oresc

-0.66

spect

-0.66

ivity

-0.63

STON

-0.61

pha

-0.61

vae

-0.59

tera

-0.57

rises

-0.57

nect

-0.57

rast

-0.56

POSITIVE LOGITS

udge

0.66

gery

0.61

hog

0.60

zzle

0.54

etz

0.53

elta

0.53

butt

0.51

 Whedon

0.51

hole

0.50

aga

0.50

Activations Density 8.218%

references to legal or judicial terms

No Comments

No Known Activations