INDEX

Explanations

technical terms related to web annotation and collaboration platforms

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

stellar

-0.89

FUL

-0.72

 diplom

-0.66

)=(

-0.65

é¾įå¥ĳå£«

-0.64

 acknow

-0.63

theless

-0.62

Dar

-0.61

 correct

-0.58

é¾įå

-0.58

POSITIVE LOGITS

rors

1.03

acts

1.02

ricks

0.98

alyses

0.97

istries

0.96

ilities

0.95

ifts

0.95

ancers

0.94

umper

0.93

enses

0.92

Activations Density 2.482%

technical terms related to web annotation and collaboration platforms

No Comments

No Known Activations