INDEX

Explanations

technical information related to programming languages, codes, protocols, and hardware specifications

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

 broader

-0.66

zens

-0.65

notations

-0.63

 however

-0.63

choes

-0.62

blance

-0.59

olulu

-0.59

 destro

-0.59

their

-0.59

yet

-0.59

POSITIVE LOGITS

->

0.97

0.89

 âľ

0.87

|--

0.83

-->

0.82

 Â·

0.80

 Received

0.80

[/

0.79

 Requires

0.79

*/

0.78

Activations Density 2.365%

technical information related to programming languages, codes, protocols, and hardware specifications

No Comments

No Known Activations