INDEX

Explanations

references to specific entities, possibly related to finance and medical practices

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_fs768-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_768

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

768

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

nod

-0.75

 soaring

-0.75

 Fletcher

-0.74

 accus

-0.74

 creep

-0.73

 bids

-0.72

 tender

-0.70

 usher

-0.70

 reve

-0.70

 crooked

-0.69

POSITIVE LOGITS

etc

1.41

Flickr

1.25

whatever

1.22

oct

1.19

index

1.19

articles

1.16

default

1.16

photos

1.15

sites

1.15

images

1.15

Activations Density 2.125%

references to specific entities, possibly related to finance and medical practices

No Comments

No Known Activations