INDEX

Explanations

demonstrations of strong criticism or condemnation

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Annotations

-0.69

lda

-0.63

intendent

-0.63

 permitting

-0.62

LER

-0.62

 Mandatory

-0.62

ramid

-0.62

 supper

-0.61

 Carbuncle

-0.61

LES

-0.61

POSITIVE LOGITS

unci

1.57

unciation

1.53

izens

1.39

ounces

1.25

igration

1.24

igrated

1.16

uclear

1.15

uded

1.13

izen

1.12

ormal

1.11

Activations Density 0.014%

demonstrations of strong criticism or condemnation

No Comments

No Known Activations