INDEX

Explanations

references to specific locations or events with historical or cultural significance

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

arians

-0.74

ional

-0.72

quit

-0.70

load

-0.69

fn

-0.66

pointers

-0.66

abel

-0.66

handedly

-0.65

edly

-0.65

ibl

-0.65

POSITIVE LOGITS

 midst

1.60

 vicinity

1.37

 aftermath

1.30

 meantime

1.30

 same

1.24

 guise

1.22

 slightest

1.16

 absence

1.13

 confines

1.12

 realm

1.10

Activations Density 2.077%

references to specific locations or events with historical or cultural significance

No Comments

No Known Activations