INDEX

Explanations

technical details and historical information about various subjects

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

 factories

-0.71

 intensive

-0.65

 relations

-0.63

 fences

-0.63

 manufact

-0.63

 intellig

-0.63

Lia

-0.62

 halfway

-0.62

 piping

-0.62

 redu

-0.62

POSITIVE LOGITS

...]

1.56

?]

1.55

âĢ¦]

1.51

sic

1.41

!]

1.38

Pg

1.37

REDACTED

1.26

actionDate

1.23

:]

1.22

1.19

Activations Density 1.540%

technical details and historical information about various subjects

No Comments

No Known Activations