INDEX

Explanations

references to awards, nominations, specific events, and achievements

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_fs1536-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_1536

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

1,536

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

 Anon

-0.79

agen

-0.79

FTP

-0.68

Alz

-0.66

hed

-0.64

folk

-0.64

STD

-0.64

omorphic

-0.63

gun

-0.63

 Mast

-0.61

POSITIVE LOGITS

 winner

1.30

 awarded

1.27

 winners

1.26

 Winners

1.22

Winner

1.18

 awards

1.17

 Award

1.14

 finalists

1.12

 medal

1.11

 nominations

1.07

Activations Density 2.649%

references to awards, nominations, specific events, and achievements

No Comments

No Known Activations