INDEX

Explanations

references to specific fictional or historical characters, locations, and concepts

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_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Anthem

-0.72

 Bleach

-0.69

 Kenobi

-0.68

wagen

-0.66

rans

-0.64

 Nirvana

-0.61

Obj

-0.60

 backs

-0.59

 internationally

-0.58

 juggling

-0.57

POSITIVE LOGITS

ASY

1.18

nerg

1.14

fficient

1.13

ighty

1.11

isner

1.08

TERN

1.07

cosystem

1.07

AST

1.07

tymology

1.07

YE

1.06

Activations Density 0.021%

references to specific fictional or historical characters, locations, and concepts

No Comments

No Known Activations