INDEX

Explanations

instances of actions or behaviors, including both positive and negative actions, with a focus on holding individuals accountable for their actions

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

bid

-0.72

anus

-0.63

ixed

-0.59

AES

-0.56

skinned

-0.56

vu

-0.56

lined

-0.55

ickets

-0.55

ANC

-0.55

ERA

-0.54

POSITIVE LOGITS

hops

0.86

 actions

0.85

 ACTIONS

0.82

uate

0.80

uations

0.79

avior

0.77

uits

0.75

uation

0.73

uated

0.71

 toward

0.70

Activations Density 11.370%

instances of actions or behaviors, including both positive and negative actions, with a focus on holding individuals accountable for their actions

No Comments

No Known Activations