INDEX

Explanations

phrases related to revelations or the conclusion of an event

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

kus

-0.74

nesota

-0.73

efer

-0.73

itude

-0.72

capital

-0.72

estine

-0.69

ciating

-0.69

eto

-0.68

pec

-0.68

ishops

-0.68

POSITIVE LOGITS

 curtains

1.31

 curtain

1.26

 veil

1.07

 shroud

0.96

 Veil

0.86

 canopy

0.82

lid

0.82

glass

0.80

 draped

0.79

 shades

0.78

Activations Density 0.031%

phrases related to revelations or the conclusion of an event

No Comments

No Known Activations