INDEX

Explanations

technical information about structures, objects, and physical properties

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

<[

-0.57

 Minor

-0.55

tid

-0.54

ointed

-0.54

 alike

-0.54

 hairc

-0.53

NB

-0.52

 Partial

-0.51

LB

-0.51

ischer

-0.51

POSITIVE LOGITS

 previously

1.26

 normally

1.25

 ordinarily

1.20

usual

1.17

 usual

1.16

 traditionally

1.13

 hitherto

1.11

 predecessors

1.03

 accustomed

1.01

 formerly

1.01

Activations Density 1.438%

technical information about structures, objects, and physical properties

No Comments

No Known Activations