INDEX

Explanations

patterns related to predicting future outcomes or probabilities

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

atin

-0.80

atri

-0.76

undy

-0.73

unal

-0.72

tha

-0.72

ansk

-0.71

ento

-0.68

untu

-0.66

gur

-0.65

gencies

-0.65

POSITIVE LOGITS

 doom

0.89

 Prediction

0.83

 predictions

0.82

 predicts

0.82

 predict

0.80

 forecasts

0.80

 prediction

0.80

pred

0.77

 accur

0.74

 Predict

0.73

Activations Density 10.886%

patterns related to predicting future outcomes or probabilities

No Comments

No Known Activations