INDEX

Explanations

phrases related to exploration or discovery

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

Data Type

torch.float32

Hook Point

blocks.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Methods

-0.84

 channels

-0.81

 establishments

-0.74

 newsletters

-0.74

 districts

-0.73

rums

-0.71

 branches

-0.70

Methods

-0.70

 sights

-0.69

 sects

-0.69

POSITIVE LOGITS

 apiece

0.87

 thereof

0.87

breaker

0.85

of

0.83

cation

0.82

pour

0.81

 consisting

0.79

onement

0.78

earance

0.78

athon

0.77

Activations Density 1.883%

phrases related to exploration or discovery

No Comments

No Known Activations