INDEX

Explanations

instances where something is discovered or located

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

idium

-0.76

puff

-0.72

panic

-0.68

assisted

-0.66

inion

-0.64

 concess

-0.62

 bailed

-0.61

rons

-0.60

 Mandatory

-0.60

rolet

-0.59

POSITIVE LOGITS

ById

1.01

 loopholes

0.93

 clues

0.92

ibility

0.83

 ways

0.82

ä½¿

0.81

AppData

0.81

 elusive

0.81

ãĤ¤ãĥĪ

0.79

 traces

0.79

Activations Density 3.619%

instances where something is discovered or located

No Comments

No Known Activations