INDEX

Explanations

words related to controversial topics, debates, and analysis

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

æĦ

-0.84

fml

-0.82

kins

-0.81

Wan

-0.79

actionDate

-0.78

uay

-0.77

incial

-0.77

Filename

-0.76

çĭ

-0.75

Tro

-0.75

POSITIVE LOGITS

consuming

0.89

ourses

0.86

 commute

0.83

 scheduling

0.82

 commuting

0.79

 reckoning

0.73

 adjourn

0.72

 continuum

0.71

 spent

0.70

theless

0.70

Activations Density 12.749%

words related to controversial topics, debates, and analysis

No Comments

No Known Activations