INDEX

Explanations

numerical patterns within parentheses representing characters or attributes of characters in a fantasy setting

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

Spec

-0.84

rights

-0.83

Rus

-0.81

HAM

-0.79

Attempts

-0.77

olson

-0.75

payers

-0.74

 composer

-0.73

pmwiki

-0.73

ãĥĥãĥī

-0.72

POSITIVE LOGITS

 elemental

0.82

 Demonic

0.76

Bai

0.76

 bushes

0.73

 Elemental

0.73

 mummy

0.72

 holy

0.71

 demonic

0.69

 Thief

0.68

 Magical

0.67

Activations Density 0.010%

numerical patterns within parentheses representing characters or attributes of characters in a fantasy setting

No Comments

No Known Activations